Re: Software RAID1 problems/questions

[Robert Dale]

On Tue, 11 Jul 2000, Tamas Acs wrote:

> One quick question, did you patch the 2.2.16 kernel with:
> http://people.redhat.com/mingo/raid-patches/raid-2.2.16-A0 ?

No, but thank you very much!!  I would like to know where you got
that information.  I snooped around and the most I could figure out
is that the patches on kernel.org stop at 2.2.11.  Even the raidtools
pointer is dead in the kernel docs.

> Tamas.
> 
> On Tue, 11 Jul 2000, Robert Dale wrote:
> 
> > 
> > Hello.
> > 
> > To start, I installed RedHat 6.1 with root on RAID1.  /boot is non-RAID.
> > It's using RedHat's 2.2.12-20 kernel which seems to have extra RAID stuff in it.
> > RAID1 is actually built as a module so it loads an initrd that has only
> > raid1.o, sh, and insmod in it.
> > 
> > I tried to upgrade to 2.2.16 by building the proper RAID support but to no
> > avail.  Basically, it does not seem to autodetect the RAID devices.  So, I read
> > md.txt and it says I need to pass md=x,x,x,x,x but it doesn't work for RAID1.
> > Well, I tried it anyway and it gets close but complains that the devices are
> > not linear or striped.  Well, duh.
> > 
> > I'm left wondering why redhat's hacks to 2.2.12 have seemingly not made it into
> > the kernel after all this time.  Am I missing some obvious or obfuscated step?
> > Are there any plans for RAID1 boot/autodetect support?
> > 
> > I don't see us moving our production systems to 2.4 before the end of the year,
> > so I don't care about 2.4's support.  If there are no other plans to do this
> > integration into 2.2 then I wouldn't mind spending a day or two patching in
> > redhat's changes.
> > 
> > Ideas?
> > 
> > For shits and giggles, here is what I have experienced so far:
> > 
> > RedHat 2.2.12 booting:
> > md driver 0.90.0 MAX_MD_DEVS=256, MAX_REAL=12
> > md.c: sizeof(mdp_super_t) = 4096
> > partition check:
> > hda: hda1 hda2 hda4 < hda5 hda6 hda7 hda8 hda9 hda10 >
> > hdc: hdc1 hdc2 hdc4 < hdc5 hdc6 hdc7 hdc8 hdc9 hdc10 >
> > autodetecting RAID arrays
> >  (read) hda5's sb offset: 524544 [events: 001f]
> >  (read) hda6's sb offset: 2097536 [events: 001f]
> >  (read) hda7's sb offset: 2097536 [events: 001f]
> >  (read) hda8's sb offset: 524544 [events: 001f]
> > .
> > .
> > .
> >  (read) hdc10's sb offset: 7307392 [events: 0008]
> > autorun ...
> > considering hdc10 ...
> > (and so on... although I don't understand why the version number is 0.90.0)
> > 
> > 
> > 2.2.16 booting:
> > md driver 0.36.6 MAX-MD_DEV=4, MAX_REAL=8
> > raid1 personality registered
> > partition check:
> > hda: hda1 hda2 hda4 < hda5 hda6 hda7 hda8 hda9 hda10 >
> > hdc: hdc1 hdc2 hdc4 < hdc5 hdc6 hdc7 hdc8 hdc9 hdc10 >
> > Oops! md0 not running, giving up!
> > hdc: hdc1 hdc2 hdc4 < hdc5 hdc6 hdc7 hdc8 hdc9 hdc10 >
> > Oops! md0 not running, giving up!
> > Bad md_map in ll_rw_block
> > EXT2-fs: unable to read superblock
> > Oops! md0 not running, giving up!
> > Bad md_map in ll_rw_block
> > isofs_read_super: bread failed, dev=09:00, iso_blknum=16, block=32
> > kernel panic: VFS: Unable to mount root fs on 09:00
> > 
> > -- 
> > Robert Dale
> > 
> >Digital Mission
> > 
> > 
> > 
> > 
> 

Software RAID1 problems/questions

[Robert Dale]

Hello.

To start, I installed RedHat 6.1 with root on RAID1.  /boot is non-RAID.
It's using RedHat's 2.2.12-20 kernel which seems to have extra RAID stuff in it.
RAID1 is actually built as a module so it loads an initrd that has only
raid1.o, sh, and insmod in it.

I tried to upgrade to 2.2.16 by building the proper RAID support but to no
avail.  Basically, it does not seem to autodetect the RAID devices.  So, I read
md.txt and it says I need to pass md=x,x,x,x,x but it doesn't work for RAID1.
Well, I tried it anyway and it gets close but complains that the devices are
not linear or striped.  Well, duh.

I'm left wondering why redhat's hacks to 2.2.12 have seemingly not made it into
the kernel after all this time.  Am I missing some obvious or obfuscated step?
Are there any plans for RAID1 boot/autodetect support?

I don't see us moving our production systems to 2.4 before the end of the year,
so I don't care about 2.4's support.  If there are no other plans to do this
integration into 2.2 then I wouldn't mind spending a day or two patching in
redhat's changes.

Ideas?

For shits and giggles, here is what I have experienced so far:

RedHat 2.2.12 booting:
md driver 0.90.0 MAX_MD_DEVS=256, MAX_REAL=12
md.c: sizeof(mdp_super_t) = 4096
partition check:
hda: hda1 hda2 hda4 < hda5 hda6 hda7 hda8 hda9 hda10 >
hdc: hdc1 hdc2 hdc4 < hdc5 hdc6 hdc7 hdc8 hdc9 hdc10 >
autodetecting RAID arrays
 (read) hda5's sb offset: 524544 [events: 001f]
 (read) hda6's sb offset: 2097536 [events: 001f]
 (read) hda7's sb offset: 2097536 [events: 001f]
 (read) hda8's sb offset: 524544 [events: 001f]
.
.
.
 (read) hdc10's sb offset: 7307392 [events: 0008]
autorun ...
considering hdc10 ...
(and so on... although I don't understand why the version number is 0.90.0)


2.2.16 booting:
md driver 0.36.6 MAX-MD_DEV=4, MAX_REAL=8
raid1 personality registered
partition check:
hda: hda1 hda2 hda4 < hda5 hda6 hda7 hda8 hda9 hda10 >
hdc: hdc1 hdc2 hdc4 < hdc5 hdc6 hdc7 hdc8 hdc9 hdc10 >
Oops! md0 not running, giving up!
hdc: hdc1 hdc2 hdc4 < hdc5 hdc6 hdc7 hdc8 hdc9 hdc10 >
Oops! md0 not running, giving up!
Bad md_map in ll_rw_block
EXT2-fs: unable to read superblock
Oops! md0 not running, giving up!
Bad md_map in ll_rw_block
isofs_read_super: bread failed, dev=09:00, iso_blknum=16, block=32
kernel panic: VFS: Unable to mount root fs on 09:00

-- 
Robert Dale

   Digital Mission

Hot Swapping disks (only somewhat on topic)

[Robert]

I am playing around with hot swapping scsi hard drives.  My machine is an
IBM netfinity 5100.  Generally I have not been able to hot swat my drives.  
For the moment, I am not using RAID, I thought it better to try a simple
ext2 file system first.  Perhaps hot swapping only works if the only
things accessed on the drives are raid areas.  Anyway, my question is how
is hot swapping supposed to work as far as linux is concerned?

This is what I find

1. If a drive is not present when the machine is booted, but is inserted
after the machine is up, it can't be mounted. (You get a message that the
device is not a valid block device.)

2. If a drive is present at boot time, and then later removed and then
later yet re-inserted, it cannot be mounted. This occurs even if the
drive was never accessed before re-inserting. (You get scsi disk error
messages.)

Is there some step I am missing? If not, I do not think "hot-swap" drives
work very well with linux, since any change in the disks requires
rebooting.  Granted I did not have to power the machine off/on, but I do
seem to have to reboot.  Perhaps I am just miss-understanding the meaning
of "hot-swap".

Best Regards,
Robert Laughlin

RE: Raid level 0 on an SMP Sparc 20

[Robert]

On Tue, 27 Jun 2000, Gregory Leblanc wrote:
> There are (or at least were in 2.2.14/15) some endian-ness issues with
> persistent superblocks.  There was a patch, and a location to download the
> patch, posted to this list a while ago, unfortunately I didn't keep that
> mail, so you'll have to check the archives.
>   Grego
> 

Thanks for this lead, I was able to find the patch at:

http://www.engin.umich.edu/caen/systems/Linux/code/patches/raid-2.2.14-B1-fix.patch

I downloaded the RedHat sources for version 2.2.16 (the latest on the
RedHat site) and the patches apply cleanly to that version.  However for
some reason I am not able to get the kernel to compile.  Here is the error
I am getting:

# make boot
gcc -Wall -Wstrict-prototypes -O2 -fomit-frame-pointer -o
scripts/split-include scripts/split-include.c
In file included from /usr/include/bits/errno.h:25,
 from /usr/include/errno.h:36,
 from scripts/split-include.c:26:
/usr/include/linux/errno.h:4: asm/errno.h: No such file or directory
make: *** [scripts/split-include] Error 1


Looking over the Makefile, this directory "asm" seems to be included in
the search path and the file is indeed in that directory.  I do not see
why the header file is not being found.  The "make depend" ran without
error.

Any assistance would be greatly appreciated.

Best Regards,
Robert Laughlin

Raid level 0 on an SMP Sparc 20

[Robert]

I am converting a Sparc-station 20 to linux, specifically RedHat 6.2.  I
have been having lots of problems with getting my RAID array to work.  I
have experience with RAID-0 and RAID-1 working fine on Intel processors.
This is my first attempt at using Linux (and RAID) on a Sparc.  I have
tried many different setups to get the RAID to work.  Some of them worked
occasionally.  All of them were either intermittant or failed to work at
all.

The three main ways I have tried to make RAID to work are:

* automatic raid, setting partition types to: fd   
When I used this method, the machine will not boot reliably.  The
errors varied.  Sometimes other (non-raid) file systems would not mount,
sometimes certain modules would not be loaded or found.

* raid startup with RedHat code in rc.sysinit.  (partitions set to 83)  
When using this technique, the boot process would usually stop, dropping
me into the maintance mode.

* raid startup with a manual command. (partitions set to 83, rc.sysinit
code modified to prevent it from starting raid)  Raid startup up usually
fails with messages I amd not used to.  For example:

command: mkraid /dev/md0
response: Killed  
after which: cat /proc/mdstat
response: Killed


Here are some details about my setup that may help...

df
Filesystem   1k-blocks  Used Available Use% Mounted on
/dev/sda4  1237672188340986460  16% /
/dev/sda115858  7142  7888  48% /boot
/dev/sda2   720784  3856680316   1% /var

cat /etc/raidtab
raiddev   /dev/md0
raid-level  0
nr-raid-disks   10
persistent-superblock   1
chunk-size  4

device  /dev/sdb1
raid-disk   0
device  /dev/sdc1
raid-disk   1
device  /dev/sdd1
raid-disk   2
device  /dev/sde1
raid-disk   3
device  /dev/sdf1
raid-disk   4
device  /dev/sdg1
raid-disk   5
device  /dev/sdh1
raid-disk   6
device  /dev/sdi1
raid-disk   7
device  /dev/sdj1
raid-disk   8
device  /dev/sdk1
raid-disk   9

free
 total   used   free sharedbuffers cached
Mem:256560  22084 234476  10516   2828  11664
-/+ buffers/cache:   7592 248968
Swap:91216  0  91216


cat /proc/cpuinfo
cpu: ROSS HyperSparc RT625 or RT626
fpu: ROSS HyperSparc combined IU/FPU
promlib: Version 3 Revision 2
prom: 2.22
type: sun4m
ncpus probed: 4
ncpus active: 4
Cpu0Bogo: 125.33
Cpu1Bogo: 125.33
Cpu2Bogo: 125.33
Cpu3Bogo: 125.33
MMU type: ROSS HyperSparc
invall: 0
invmm: 0
invrnge : 0
invpg   : 0
contexts: 4096
CPU0: online
CPU1: online
CPU2: online
CPU3: online
  
typical fdisk partition setup:
Disk /dev/sdc (Sun disk label): 67 heads, 62 sectors, 1007 cylinders
Units = cylinders of 4154 * 512 bytes

   Device FlagStart   EndBlocks   Id  System
/dev/sdc1 1  1007   2089462   83  Linux native
/dev/sdc3 0  1007   20915395  Whole disk


Any assistance would be greatly appreciated, I have been banging my head
over this for about 10 days now...:(

Best Regards,
Robert Laughlin

Re: resyncing takes forever again

[Robert Stuart]

Brian Grossman wrote:
> 
> > I found the message at the bottom on the mailing list archives...
> >
> > I'd like to register a "me too" with this problem.  Was there any
> > further discussion?
> >
> > I've got a 2.2.14 kernel with the Mingo patch 2.2.14-B1, and I get the
> > same problem.  I get no scsi errors, no other kernel messages after the
> > raid reconstruction starts, and the completion time slowly climbs.  This
> > has happened on a couple of machines now.
> 
> Does something like 'echo 5120 > /proc/sys/dev/md/speed-limit' make a
> difference?
> 

I've got a mirror in this state now, and have tried the above with no
difference.  I should have pointed out that the recover percentage
doesn't change.  Brian, I'm not sure whether you were asking me if this
makes a difference now or later...

My mdstat from this morning:

# cat /proc/mdstat 
Personalities : [raid1] 
read_ahead 1024 sectors
md0 : active raid1 sdd2[1] sda2[0] 1024960 blocks [2/2] [UU]
md3 : active raid1 sdd3[1] sda3[0] 7846848 blocks [2/2] [UU]
md2 : active raid1 sde1[2] sdb1[1] 35881024 blocks [2/1] [_U]
recovery=28% finish=6126.7min
md1 : active raid1 sdf1[1] sdc1[0] 17775808 blocks [2/2] [UU]
unused devices: 


Thanks
-- 
Robert Stuart
Systems Administrator
ph: 3864 0364

resyncing takes forever again

[Robert Stuart]

Hi,

Note: Could people email me directly please as I am not on the list.

I found the message at the bottom on the mailing list archives...

I'd like to register a "me too" with this problem.  Was there any
further discussion?

I've got a 2.2.14 kernel with the Mingo patch 2.2.14-B1, and I get the
same problem.  I get no scsi errors, no other kernel messages after the
raid reconstruction starts, and the completion time slowly climbs.  This
has happened on a couple of machines now.  

There is the possibility that the drive is faulty (its hard to be sure
as there were a few SCSI bus errors that affected a couple of other
devices too before it got removed from the mirror).

#uname -a
Linux fsx.bq.edu.au 2.2.14-fsx-2421 #1 Fri Apr 21 10:26:45 EST
2000 i686 unknown

#cat /proc/mdstat
Personalities : [raid1] 
read_ahead 1024 sectors
md0 : active raid1 sdd2[1] sda2[0] 1024960 blocks [2/2] [UU]
md3 : active raid1 sdd3[1] sda3[0] 7846848 blocks [2/2] [UU]
md2 : active raid1 sde1[2] sdb1[1] 35881024 blocks [2/1] [_U]
recovery=28% finish=3168.1min
md1 : active raid1 sdf1[1] sdc1[0] 17775808 blocks [2/2] [UU]
unused devices: 

Also can I suggest that whoever maintains this list, creates a
signature/footer to the bottom of every email that contains the mailing
list archive address and the faq/howto?

Any help pointers etc would be most appreciated.


-
begin previous message:

On Friday, May 26, 2000, Thomas Gebhardt said:



> /proc/mdstat says that the new disk is resyncing but there is no
> progress at all. There is only 1% of the mirror in sync now and
> the estimated time of completion grows continuously. There is
> a load average of about 1.0 caused by this process.
> 
I have seen the same thing. 0% reported progress, 
time estimate runs away. This after doing 'raidsetfaulty', 
'raidhotremove', and 'raidhotadd' on an OK disk partition - 
I was just doing some testing.

> After several trys I gave up to resync the mirror and started
> the server with the incomplete RAID 1 Array :-(
> 
For me, the only way to stop this was to reboot (raidstop 
just hung). Even reboot was not clean as the offending array 
wouldn't stop during shut down. Had to press reset - causing 
long re-sync of other arrays after restart.

> Environment:
> Kernel is 2.2.10 with new-raid-style patches applied, auto-detection
> activated and a persistent superblock.
> 
I saw this with clean 2.2.14, mingo's 2.2.14-B1 patch, 
RAIDTOOLS-19990824-0_90_TAR.GZ, autodetect enabled raid 
in the kernel (not a module).

> Any hints?
> 
> Thanks, Thomas


-- 
Robert Stuart
Systems Administrator
ph: 3864 0364

Re: Problem setting up a raid0 with raidtools-0.90-6 and redhat-6.2

[Robert]

On Sat, 20 May 2000, Harry Zink wrote:
> on 5/20/00 9:11 AM, Robert at [EMAIL PROTECTED] wrote:
> 
> > I am probably doing something silly, put when applying the patch, lots
> > of the hunks seem to get rejected.  Any ideas?
> 
> Did you:
> 
> patch -p0  
> ??
> 
> Did you apply them to a new, downloaded kernel?
> 
> Harry
> 

Thanks for replying, Harry.  Turns out what was wrong, was that the patch
file had CR (carriage return) chars in it.  I downloaded it into Windows
and the locally ftp'd it to Linux.  The first transfer was in text mode, 
the second was in binary.  That is how the CRs got in there.  It seems
that patch can't deal with CRs.

Best Regards,
Robert Laughlin

Re: Problem setting up a raid0 with raidtools-0.90-6 and redhat-6.2

[Robert]

> On Tue, 16 May 2000, Harry Zink wrote:
> > You can get the patches at:
> > 
> > http://people.redhat.com/mingo/raid-patches/
> > 
> > Apply them, re-compile your kernel, and this should work.
> > 
> > Harry
> > 

Is there any other place to get these patches? Neither the one for 2.2.14
nor the one for 2.2.15 on this site seems to work for me.  Both reject
*lots* of hunks (but not all).

Best Regards,
Robert Laughlin

Re: Problem setting up a raid0 with raidtools-0.90-6 and redhat-6.2

[Robert]

On Tue, 16 May 2000, Harry Zink wrote:
> You can get the patches at:
> 
> http://people.redhat.com/mingo/raid-patches/
> 
> Apply them, re-compile your kernel, and this should work.
> 
> Harry
> 

http://www.kernel.org/pub/linux/kernel/v2.2/linux-2.2.1503-Mat-2000 17:22

 does not seem to be able to be patched using

http://people.redhat.com/mingo/raid-patches/raid-2.2.15-A0  19-May-2000 18:07


I am probably doing something silly, put when applying the patch, lots
of the hunks seem to get rejected.  Any ideas?

Best Regards,
Robert Laughlin

Re: /etc/raidtab lost

[Robert]

I am guessing that my problem are the files /etc/md0 and /etc/md1 which
disappeared when I lost my /dev directory.  Is there any way to recreate
these short of rebuilding and therefor destroying the underlining data?
Every raid command I try to run gives me the same error message:

"cannot determine md version: No such file or directory"

It certainly would be better if the error messages where more descriptive.


On Tue, 16 May 2000, Robert wrote:

> I should probably add that the RAID areas are not system areas.  They
> contain data only, not any operating system files.
> 
> 
> On Mon, 15 May 2000, Robert wrote:
> 
> > Not sure why, but last night my /etc and /var and /bin directories
> > dissappeared.  Not sure if it was a bug, or some hacker who found his way
> > into my machine.  Needless to say, I had to reinstall most everything.
> > 
> > One of the files I lost was the most up to date copy of /etc/raidtab.  I
> > have used a backup and tried to remember the changes since then, but so
> > far no luck.  Is there anyway for me to get my raid arrays running again
> > now?
> > 
> > details:  kernal 2.2.13  with raid0145-19990824-2.2.11
> > I do not have the partitions set to FD, I am not using auto-detection.
> > 
> > --raidtab
> > I am sure the devices and the order of them is correct. I am not sure
> > about the chuck-size.
> > -
> > raiddev /dev/md0
> > raid-level  0
> > nr-raid-disks   4   
> > nr-spare-disks  0
> > persistent-superblock   1
> > chunk-size  4
> > 
> > device  /dev/hda5
> > raid-disk   0
> > device  /dev/hdc1
> > raid-disk   1
> > device  /dev/hdd1
> > raid-disk   2
> > device  /dev/hde1
> > raid-disk   3
> > 
> > raiddev /dev/md1
> > raid-level  1
> > nr-raid-disks   2   
> > nr-spare-disks  0
> > persistent-superblock   1
> > chunk-size  4
> > 
> > device  /dev/hdc7
> > raid-disk   0
> > device  /dev/hdd6
> > raid-disk   1
> > -
> > # raidstart /dev/md0
> > cannot determine md version: No such file or directory
> > # raidstart /dev/md1
> > cannot determine md version: No such file or directory
> > # cat /proc/mdstat
> > Personalities : [linear] [raid0] [raid1]
> > read_ahead not set
> > unused devices: 
> > 
> > Best Regards,
> > Robert Laughlin
> > 
> > 
> 
> Best Regards,
> Warm Regards,
> Hugs,
> Robert L.
> Robert Laughlin
> 
> 

Best Regards,
Warm Regards,
Hugs,
Robert L.
Robert Laughlin

Re: /etc/raidtab lost

[Robert]

I should probably add that the RAID areas are not system areas.  They
contain data only, not any operating system files.


On Mon, 15 May 2000, Robert wrote:

> Not sure why, but last night my /etc and /var and /bin directories
> dissappeared.  Not sure if it was a bug, or some hacker who found his way
> into my machine.  Needless to say, I had to reinstall most everything.
> 
> One of the files I lost was the most up to date copy of /etc/raidtab.  I
> have used a backup and tried to remember the changes since then, but so
> far no luck.  Is there anyway for me to get my raid arrays running again
> now?
> 
> details:  kernal 2.2.13  with raid0145-19990824-2.2.11
> I do not have the partitions set to FD, I am not using auto-detection.
> 
> --raidtab
> I am sure the devices and the order of them is correct. I am not sure
> about the chuck-size.
> -
> raiddev /dev/md0
>   raid-level  0
>   nr-raid-disks   4   
>   nr-spare-disks  0
> persistent-superblock   1
> chunk-size  4
> 
>   device  /dev/hda5
>   raid-disk   0
>   device  /dev/hdc1
>   raid-disk   1
>   device  /dev/hdd1
>   raid-disk   2
>   device  /dev/hde1
>   raid-disk   3
> 
> raiddev /dev/md1
>   raid-level  1
>   nr-raid-disks   2   
>   nr-spare-disks  0
> persistent-superblock   1
> chunk-size  4
> 
>   device  /dev/hdc7
>   raid-disk   0
>   device  /dev/hdd6
>   raid-disk   1
> -
> # raidstart /dev/md0
> cannot determine md version: No such file or directory
> # raidstart /dev/md1
> cannot determine md version: No such file or directory
> # cat /proc/mdstat
> Personalities : [linear] [raid0] [raid1]
> read_ahead not set
> unused devices: 
> 
> Best Regards,
> Robert Laughlin
> 
> 

Best Regards,
Warm Regards,
Hugs,
Robert L.
Robert Laughlin

/etc/raidtab lost

[Robert]

Not sure why, but last night my /etc and /var and /bin directories
dissappeared.  Not sure if it was a bug, or some hacker who found his way
into my machine.  Needless to say, I had to reinstall most everything.

One of the files I lost was the most up to date copy of /etc/raidtab.  I
have used a backup and tried to remember the changes since then, but so
far no luck.  Is there anyway for me to get my raid arrays running again
now?

details:  kernal 2.2.13  with raid0145-19990824-2.2.11
I do not have the partitions set to FD, I am not using auto-detection.

--raidtab
I am sure the devices and the order of them is correct. I am not sure
about the chuck-size.
-
raiddev /dev/md0
raid-level  0
nr-raid-disks   4   
nr-spare-disks  0
persistent-superblock   1
chunk-size  4

device  /dev/hda5
raid-disk   0
device  /dev/hdc1
raid-disk   1
device  /dev/hdd1
raid-disk   2
device  /dev/hde1
raid-disk   3

raiddev /dev/md1
raid-level  1
nr-raid-disks   2   
nr-spare-disks  0
persistent-superblock   1
chunk-size  4

device  /dev/hdc7
raid-disk   0
device  /dev/hdd6
raid-disk   1
-
# raidstart /dev/md0
cannot determine md version: No such file or directory
# raidstart /dev/md1
cannot determine md version: No such file or directory
# cat /proc/mdstat
Personalities : [linear] [raid0] [raid1]
read_ahead not set
unused devices: 

Best Regards,
Robert Laughlin

dump segmentation fault on RAID 0+1 partition

[Gerrish, Robert]

I am running Red Hat 6.1 on this particular computer which appears
to have raidtools v0.90, etc . . . all the latest working patches.
 
I had two mirrored partitions that were 8 & 6GB and we needed a 14GB
partition.  Rather than repartitioning the disks (and having to 
reinstalling the system), I concatenated the partitions and then 
remirrored (from raid 1 ro raid 1+0).  I did a mkfs on the file system, 
mounted it and it appears to work OK.  Since than, I have not been able 
to backup the partition with dump.  I can use tar and it appears to work.

The output of the dump command is:
   DUMP: Date of this level 0 dump: Tue Apr 18 14:48:51 2000 
   DUMP: Date of last level 0 dump: the epoch
   DUMP: Dumping /dev/md8 (/apps) to /dev/null   
   DUMP: mapping (Pass I) [regular files]
   DUMP: mapping (Pass II) [directories] 
   DUMP: estimated 260613 tape blocks on 0.01 tape(s).   
   DUMP: Volume 1 started at: Tue Apr 18 14:50:02 2000   
   DUMP: dumping (Pass III) [directories]
   DUMP: SIGSEGV: ABORTING!  
   DUMP: SIGSEGV: ABORTING!  
   Segmentation fault  
 
The original devices were md6 as a mirror and md7 as a mirror.  
I did a raidstop, on the old devices, reconfigured raidtab and 
did a mkraid on md6 and md7 then on md8 and then a mkfs on md8.
 
My /etc/raidtab for those devices is:
   raiddev /dev/md6   
   raid-level  0  
   nr-raid-disks   2  
   chunk-size  64k
   persistent-superblock   1  
   #nr-spare-disks 0  
   device  /dev/hda5  
   raid-disk 0
   #device /dev/hdc1  
   device  /dev/hda12 
   raid-disk 1
   raiddev /dev/md7   
   raid-level  0  
   nr-raid-disks   2  
   chunk-size  64k
   persistent-superblock   1  
   #nr-spare-disks 0  
   device  /dev/hdc1  
   #device /dev/hda12 
   raid-disk 0
   device  /dev/hdc5  
   raid-disk 1
   raiddev /dev/md8   
   raid-level  1  
   nr-raid-disks   2  
   chunk-size  64k
   persistent-superblock   1  
   #nr-spare-disks 0  
   device  /dev/md6   
   raid-disk 0
   device  /dev/md7   
   raid-disk 1
 
Is there something I am overlooking?  Would I be better off 
reloading the system?
 
TIA,

Bob Gerrish
Unix Systems Administrator
Trim Systems, LLC
Seattle, WA
[EMAIL PROTECTED]

Software RAID1

[Robert Hélie]

Hi,

I have just finished configuring a RAID 1 array. Since my system was already
installed I used the "failed-disk" approach.

When I tried rebooting the system Linux told me that the
"persisten-superblock" was not the right size.

Here are the steps I followed:

RAIDTAB file:
raiddev /dev/md0
raid-level  1
nr-raid-disks   2
nr-spare-disks  0
chunk-size  4
persistent-superblock   1
device  /dev/sdb1
raid-disk   0
device  /dev/sda1
raid-disk   1
failed-disk 1

sdb1 is a ~17Gb partion of type "fd"
I also have sdb2 as a linux swap (~128M)

Is it necessary to have a linux swap when using a RAID? Could that be the
cause of the problem (I noticed that the reported difference in the
"persistent-superblock" was small)?

Your help on that will be greatly appreciated.

Thanks,


Robert Hélie

Re: Swap on Raid1 : safe on EIDE disks ? =no hangs ?

[Robert Dahlem]

Benno,

On Fri, 07 Jan 2000 13:07:16 +0100, Benno Senoner wrote:

>I was wondering if this procedure (swapping on Raid1 + waiting for the resync),
>is safe on an IDE only system ?

>I don't need hot-swapping, the only thing I need is that if one disk dies,
>swapping will not take down the disk (hangs like in the SCSI case etc).
>
>If this is safe I can easily live with the "wait for resync" issue.

Try this: change "swap" to "raidswap"  in /etc/fstab, then merge this procedure into 
an 
early position in your startup procedures:

>
#!/bin/sh

FSTAB=/etc/fstab

usage()
{
echo "usage: $0 start" >&2
exit 1
}

read_fstab()
{
awk ' {
if(NF && substr($1,1,1) != "#" && $3 == "raidswap")
print $1
} ' $FSTAB
}

check_raiddev()
{
raiddev=`basename $1`
status=`cat /proc/mdstat | awk '
/^'$raiddev' : / {
if($NF ~ /\[U*\]/)
print "OK"
else
print "NOK"
}
'`
case "$status" in
OK) return 0 ;;
*)  return 1 ;;
esac
}


case "$1" in
start)
read_fstab | while read device ; do
if check_raiddev $device ; then
echo "Start swapping to $device"
swapon $device
else
echo "*** Not swapping to $device: not in sync"
fi
done
;;
stop)
# Just to prevent false usage alarms
;;
*)
usage
;;
esac
>

Your system will probably give some alarms at startup ("fs type raidswap not in 
kernel" 
or so), but just don't mind.

It's a quick hack (no check for swapping priority et al), but works fine for me.

Regards,
Robert

Re: WARNING: raid for kernel 2.2.11 used with 2.2.14 panics

[Robert Dahlem]

Ingo,

On Thu, 6 Jan 2000 14:14:07 +0100 (CET), Ingo Molnar wrote:

>> I just wanted to warn everybody not to use raid0145-19990824-2.2.11
>> together with kernel 2.2.14: at least in my configuration (two IDE
>> drives with RAID-1, root on /dev/mdx) the kernel panics with "B_FREE
>> inserted into queues" at boot time.
>
>this should be fixed in:
>
>   http://www.redhat.com/~mingo/raid-2.2.14-B1
>
>let me know if you still have any problem. The problem outlined by
>Andrea's patch (which reverses a patch of mine) is solved as well.

Thank you! Indeed this fixes the kernel panic problem.

I already tried the combinations (losing disk 1, recovering from disk 2 and vice 
versa) 
and it runs perfectly smooth.

Regards,
Robert

WARNING: raid for kernel 2.2.11 used with 2.2.14 panics

[Robert Dahlem]

Hi,

I just wanted to warn everybody not to use raid0145-19990824-2.2.11 together with 
kernel 
2.2.14: at least in my configuration (two IDE drives with RAID-1, root on /dev/mdx) 
the 
kernel panics with "B_FREE inserted into queues" at boot time.

This seems to be some kind of a known problem, have a look at:

http://www.deja.com/=dnc/getdoc.xp?AN=557352317

The main statement there reads like this:

So basically in both cases the raid code is destroying the buffer data
(even if the buffer it's in-use!). So the above snapshot of code is very
buggy.

This posting also includes a patch, but seriously: I'm not the man to try a patch 
which 
seems to remove a whole function. :-)

Regards,
Robert

X windows

[Robert]

I am trying to use the X windows systems for the first time.  I am using
the NCD package on my PC as the server.  I find if I run individual
programs like xterm on my linux box, it works just fine, openning a window
on my pc.  

What I was hoping to be able to do, was to see the entire
gnome/enlightment desktop within a window on my PC.  So far I am not able
to make that work.  Apparnently, these windows managers are setup so only
one can run at a time.  And since there is one running on my linux
(attached) monitor, I can't get another one to run over X to my PC.

Am I trying to do the impossible, or is there a trick to making this work?
Also, does anyone have a recommendation for a MS Windows Xserver, or is
NCD the best choice?

Best Regards,
Robert Laughlin

Re: disks lose label at reboot, killing md device with "bad magic". Help!

[Robert Dubinski]

Thank-you.  This solved the problem.

I just found a very old FAQ that lists this problem.  It's
specific to SPARCs, where Sun writes the diskinfo in the first
1K of the disk, and raidtools then overwrites that.

I think this should get added to the "Pitfalls" section of the
newer HOWTO at ostenfeld.dk.  


-Robb







On Tue, Dec 21, 1999 at 01:28:39AM +, [EMAIL PROTECTED] wrote:
> try starting all your partitions at cyl 1 instead of 0. (ie- leave first cyl
> unused)
> 
> allan
> 
> 
> Robert Dubinski <[EMAIL PROTECTED]> said: 
> 
> > 
> > Hi,
> > 
> > I'm trying to set up a RAID-0 device over three disks in a Sparc
> > Storage Array with RedHat 6.1/SPARC, but am having a problem.  I
> > searched the archives of this list and couldn't find a solution or
> > a thread on this particular problem.
> > 
> > mkraid works fine, as well as the formatting and mounting of my
> > md0 device.  I can raidstop and raidstart the device all I
> > want, and /proc/mdstat happily reports the devices existence.  
> > >From what I can see, it works fine...until I reboot:
> > 
> > At a reboot, when the disks are checked, the latter two in the
> > RAID complain about "bad magic".  The md0 device doesn't come
> > online, and when I go to fdisk the latter two disks...the
> > disklabel and partition info have vanished.
> > 
> > I tried a number of things, such as toggling auto-detection  and
> > removing the persistent-superblock mode, but to no avail: I always
> > get the "bad magic" error upon reboot.
> > 
> > First, my raidtab:
> > 
> > #
> > raiddev /dev/md0
> > raid-level 0
> > nr-raid-disks 3
> > persistent-superblock 1
> > chunk-size 8
> > device /dev/sdc5
> > raid-disk 0
> > device /dev/sdm5
> > raid-disk 1
> > device /dev/sdw5
> > raid-disk 2
> > #
> > 
> > My partition table format:
> > 
> > Disk /dev/sdc (Sun disk label): 19 heads, 80 sectors, 2733 cylinders
> > Units = cylinders of 1520 * 512 bytes
> > 
> >Device FlagStart   EndBlocks   Id  System
> > /dev/sdc3  u  0  2733   20770805  Whole disk
> > /dev/sdc5 0  2733   2077080   83  Linux native
> > 
> > (when I toggled autodetection, I changed it to fd as the HOWTO
> > mentions).
> > 
> > And lastly, the 'dmesg' upon bootup:
> > 
> > Access  ANSI SCSI revision: 02
> > Detected scsi disk sdn at scsi1, channel 3, id 2, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
> >   Type:   Direct-Access  ANSI SCSI revision: 02
> > Detected scsi disk sdo at scsi1, channel 3, id 3, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
> >   Type:   Direct-Access  ANSI SCSI revision: 02
> > Detected scsi disk sdp at scsi1, channel 3, id 4, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0420
> >   Type:   Direct-Access  ANSI SCSI revision: 02
> > Detected scsi disk sdq at scsi1, channel 4, id 0, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
> >   Type:   Direct-Access  ANSI SCSI revision: 02
> > Detected scsi disk sdr at scsi1, channel 4, id 1, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
> >   Type:   Direct-Access  ANSI SCSI revision: 02
> > Detected scsi disk sds at scsi1, channel 4, id 2, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
> >   Type:   Direct-Access  ANSI SCSI revision: 02
> > Detected scsi disk sdt at scsi1, channel 4, id 3, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
> >   Type:   Direct-Access  ANSI SCSI revision: 02
> > Detected scsi disk sdu at scsi1, channel 4, id 4, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
> >   Type:   Direct-Access  ANSI SCSI revision: 02
> > Detected scsi disk sdv at scsi1, channel 5, id 0, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
> >   Type:   Direct-Access  ANSI SCSI revision: 02
> > Detected scsi disk sdw at scsi1, channel 5, id 1, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
> >   Type:   Direct-Access  ANSI SCSI revision: 02
> > Detected scsi disk sdx at scsi1, channel 5, id 2, lun 0
> >   Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
> >   Type:   Direct-Access 

disks lose label at reboot, killing md device with "bad magic". Help!

[Robert Dubinski]

Hi,

I'm trying to set up a RAID-0 device over three disks in a Sparc
Storage Array with RedHat 6.1/SPARC, but am having a problem.  I
searched the archives of this list and couldn't find a solution or
a thread on this particular problem.

mkraid works fine, as well as the formatting and mounting of my
md0 device.  I can raidstop and raidstart the device all I
want, and /proc/mdstat happily reports the devices existence.  
>From what I can see, it works fine...until I reboot:

At a reboot, when the disks are checked, the latter two in the
RAID complain about "bad magic".  The md0 device doesn't come
online, and when I go to fdisk the latter two disks...the
disklabel and partition info have vanished.

I tried a number of things, such as toggling auto-detection  and
removing the persistent-superblock mode, but to no avail: I always
get the "bad magic" error upon reboot.

First, my raidtab:

#
raiddev /dev/md0
raid-level 0
nr-raid-disks 3
persistent-superblock 1
chunk-size 8
device /dev/sdc5
raid-disk 0
device /dev/sdm5
raid-disk 1
device /dev/sdw5
raid-disk 2
#

My partition table format:

Disk /dev/sdc (Sun disk label): 19 heads, 80 sectors, 2733 cylinders
Units = cylinders of 1520 * 512 bytes

   Device FlagStart   EndBlocks   Id  System
/dev/sdc3  u  0  2733   20770805  Whole disk
/dev/sdc5 0  2733   2077080   83  Linux native

(when I toggled autodetection, I changed it to fd as the HOWTO
mentions).

And lastly, the 'dmesg' upon bootup:

Access  ANSI SCSI revision: 02
Detected scsi disk sdn at scsi1, channel 3, id 2, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdo at scsi1, channel 3, id 3, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdp at scsi1, channel 3, id 4, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0420
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdq at scsi1, channel 4, id 0, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdr at scsi1, channel 4, id 1, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sds at scsi1, channel 4, id 2, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdt at scsi1, channel 4, id 3, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdu at scsi1, channel 4, id 4, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdv at scsi1, channel 5, id 0, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdw at scsi1, channel 5, id 1, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdx at scsi1, channel 5, id 2, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdy at scsi1, channel 5, id 3, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdz at scsi1, channel 5, id 4, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdaa at scsi1, channel 6, id 0, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdab at scsi1, channel 6, id 1, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdac at scsi1, channel 6, id 2, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdad at scsi1, channel 6, id 3, lun 0
  Vendor: SEAGATE   Model: ST32550W SUN2.1G  Rev: 0418
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdae at scsi1, channel 6, id 4, lun 0
scsi : detected 1 SCSI cdrom 31 SCSI disks total.
esp0: target 6 asynchronous
Uniform CDROM driver Revision: 2.56
esp0: target 0 [period 100ns offset 15 20.00MHz FAST-WIDE SCSI-II]
SCSI device sda: hdwr sector= 512 bytes. Sectors= 4194995 [2048 MB] [2.0 GB]
SCSI device sdb: hdwr sector= 512 bytes. Sectors= 4194995 [2048 MB] [2.0 GB]

Re: enlarging a RAID-0

[Robert]

I am happy to report that raidreconf worked perfectly for me.  I needed to
enlarge a RAID-0 array from 43 gig (with 3 disks) to 64 gig (on 4 disks).
Before conversion the array was about 58% in use with actual data.  The
enlargement run took 29 hours to complete.  Following conversion, e2fsck
reported as follows:

> /sbin/e2fsck /dev/md0
e2fsck 1.14, 9-Jan-1999 for EXT2 FS 0.5b, 95/08/09
/dev/md0: clean, 5481/11522048 files, 26546130/46086080 blocks

Thank you, Jakob for your great reconfiguration program.  While 29 hours
is a very long time...:)  It would have taken much longer to reload
everything from the backups if I had started over from scratch.  I still
need to run ext2resize to enlarge the file system, I hope that goes as
well!

Best Regards,
Robert Laughlin

On Wed, 15 Dec 1999, Jakob Østergaard wrote:
> But anyways, I'm very happy you decided to give it a go.  If it ends
> up actually working for you too, then maybe my C hackery isn't too far
> off after all ;)

HPT366 boot problem

[Robert]

Here is the patch I used that allows my system to boot. If there is
anything more I can contribute to help create a better fix, please let me
know.  Without this patch, my system locks up on boot whenever I put a
drive on the ide3 or ide4 channels.  With this patch installed, the system
will boot successfully and I can reliably use the ide3 and ide4 attached
drives.  Andre, I am assuming your latest patch to 2.2.14pre12 does not
address this.


--- linux+raid+ide/drivers/block/hpt366.c   Tue Dec 14 12:50:59 1999
+++ linux/drivers/block/hpt366.cTue Dec 14 13:37:46 1999
@@ -192,7 +192,7 @@
reg2 &= ~0x8000;
 
pci_write_config_dword(HWIF(drive)->pci_dev, regtime, reg2);
-   err = ide_config_drive_speed(drive, speed);
+   err = 0;   /*   err = ide_config_drive_speed(drive, speed); */  
 
 #if HPT366_DEBUG_DRIVE_INFO
printk("%s: speed=0x%02x(%s), drive%d, old=0x%08x, new=0x%08x, err=0x%04x\n",




Best Regards,
Robert Laughlin

On Wed, 15 Dec 1999, Andre Hedrick wrote:
> Drop the latest stuff from
> ftp://ftp.*.kernel.org/pub/linux/kernel/people/hedrick/
> and see if that helps.

Re: Performance?

[Robert]

I was finally able to get my HPT366 channels to work! But I do not feel
all that good about how I did it.  Andre, if you are inclinded to look
into this, I hope the enclosed information is helpful to you.  If you want
any additional information, just let me know.  Please note that at no time
was I trying to boot off IDE3 or IDE4, so that was never an issue for me.

I am using an ABIT BE6 motherboard.  I have upgraded it to the "PL"
version of the BIOS.  This version of BIOS includes version 1.21 of the
HPT366 drivers (the same version as the "NJ" bios for board BP6.) My
kernel is 2.2.13 with patches: raid0145-19990824-2.2.11 and
ide.2.2.13.1999.patch applied. Before the patch I made, I was able to
pin down where the system was hanging, by adding "printk()" calls, and
seeing where things stopped. Below is where the problem seems to be
occurring:

  hpt366_tune_chiptset()
-calls-
  ide_config_drive_speed()  
-has this statement-
  if (IDE_CONTROL_REG)
  OUT_BYTE(drive->ctl | 2, IDE_CONTROL_REG); 

during/after the "OUT_BYTE" operation, the system hangs. "OUT_BYTE" is an
inline macro that sends a byte directly to the hardware, and I have
no clue why sending this particular byte is hanging the system.

My solution was to comment out the call to ide_config_drive_speed() within
hpt366_tune_chiptset() (setting err=0), and let the BIOS driver set
the speed. The new BIOS driver has a screen that lets you set the transfer
mode.  I set it to UDMA 3.

Once I made this change, I was able to boot, run cfdisk to create a
partition, and mke2fs to build a file system, then mount it. I will be
stress testing the drive over the next few days.  But I am much farther
along than I was before.  I assume the BE6 motherboard will be
increasingly popular for linux use, so I hope a better fix will be
eventually forthcoming.

I want to thank everyone who made suggestions and offered
encouragement.

Best Regards,
Robert Laughlin



On Fri, 10 Dec 1999, Tim Niemueller wrote:
> Robert wrote:
> > 
> > Tim, I have been trying for months to get the ATA66 channels to
> > work on an ABIT BE6 motherboard with no success.  At the moment, despite
> > trying numerous kernal versions, numerous patch combinations and most
> > recently two different bios version, my system still locks up on boot when
> > a drive is attached to either ide3 or ide4.  I would be very leary of
> > HPT366 at this stage.
> 
> But the Component dbase (cdb.suse.de, perhaps there is also an
> international version under cdb.suse.com, I am not using SuSE) of SuSE
> says, that the board and the board are supported by kernel from 2.2.13.
> Have a look for "Motherboards, Abit". It will show you the results.

Re: Performance?

[Robert]

Tim, I have been trying for months to get the ATA66 channels to
work on an ABIT BE6 motherboard with no success.  At the moment, despite
trying numerous kernal versions, numerous patch combinations and most
recently two different bios version, my system still locks up on boot when
a drive is attached to either ide3 or ide4.  I would be very leary of
HPT366 at this stage.

Best Regards,
Robert Laughlin

On Thu, 9 Dec 1999, Tim Niemueller wrote:

> Hi there,
> 
> I want to build up an array of four IBM DJNA 15 GB harddisks on an Abit
> PE6 with ATA/66 Controller. The array should be RAID-5, what so you know
> about the performance? I mean in general not only for this specific
> constellation? Is the source stable and usable for production?
> 
>   Tim
> 
> -- 
> 
>Tim Niemueller <[EMAIL PROTECTED]> www.niemueller.de
> 
> --> PGP-ID and Fingerprint < Get on Homepage! 
>  (RSA 0x727B2BB7) 4244 5CBB 67CC 7DC4 71B2 6EB7 149F 2585
> 

Re: RAID0 performance odditity

[Robert]

Thanks for writing back, Michael.  Yes it complied cleanly, and I have a
short script that copies over the new kernal and runs lilo for me, so I
don't have to remember all the steps...:)

I have written Andre a number of times providing him with details about my
problem, I have also added some additional debug messages to help pin down
where the lockup is occurring, but no joy so far.  

My guess is that the HTP366.c routines works for the ABIT BP6 board but
not for the ABIT BE6 board.  All of the other people who have written
indicating some success with HTP366 have the BP6 board.  I believe that
the BP6 board was the one Andre used when developing the code. One person
wrote saying the BP6 uses "NJ" bios, on the last boot I watched carefully
and my bios says it's "Award Modular 4.51PG" bios.

Best Regards,
Robert Laughlin

On Thu, 9 Dec 1999, Michael Trainor wrote:

> Interesting.  I never had lockups during boot, only during heavy IDE load.
> Just a stupid question:  you did make sure that the change was cleanly
> compiled in and installed and all?  I assume you probably did, but I've
> missed steps before when not really watching what I was doing and sat
> scratching my head for a while trying to find a problem that didn't really
> exist (forget to run lilo or something stupid).
> 
> I emailed Andre (the IDE maintainer) yesterday, and he indicated that my
> problem stemmed mostly (or entirely) from the Maxtor drives.  Apparently the
> HPT has little tolerance for variation in the ATA66 spec and the Maxtor
> drives don't adhere too well and they clash.  I guess Maxtor is unwilling to
> admit the problem with their firmware, so the problem continues.
> 
> Mike

Re: RAID0 performance odditity

[Robert]

T just tried making the same change on my system to see if it would help
me.  But the symptons stayed the same.  If a drive is attached to IDE3 or
IDE4 channels, the system locks up during bootup. One difference I am
using the BE-6 motherboard.

Best Regards,
Robert Laughlin

On Wed, 8 Dec 1999, Michael Trainor wrote:
> In any event, I ended up browsing through the code (experienced programmer,
> linux kernel newbie) and found that I could disable Ultra/66 by changing
> #DEFINE HPT366_ALLOW_ATA66_4 1 to #DEFINE HPT366_ALLOW_ATA66_4 0

HPT366 prevents boot on Abit BE-6 (probably not raid related)

[Robert]

Upon further investigation, I added some debug printk() statements.
Assuming buffering is not confusing the issue, it looks like the machine
is calling this subroutine, but the subroutine never returns;

ide_config_drive_speed(drive, speed);

The parms are set to these values when the call is made:

   drive->name = hde
   speed = 0x44

The subroutine ide_config_drive_speed() is part of ide.c

Best Regards,
Robert Laughlin

-- Forwarded message --
Date: Thu, 2 Dec 1999 09:04:10 -0500 (EST)
From: Robert <[EMAIL PROTECTED]>
To: Tim Moore <[EMAIL PROTECTED]>
Cc: raid <[EMAIL PROTECTED]>, Andre Hedrick <[EMAIL PROTECTED]>
Subject: HPT366 prevents boot on Abit BE-6 (probably not raid related)

I am still trying to get my HPT366 channels to work on my ABIT BE6
motherboard.  I really appreciate the emails I have gotten to date, but
there is still no joy.  Following in the footsteps of Tim Moore, I
upgraded to 2.2.13 with patches raid0145-19990824-2.2.11, and
ide.2.2.13.1999.patch applied.  This has made no difference.  I
noticed there is a debug option in hpt366.c, namely 
HPT366_DEBUG_DRIVE_INFO
So I turned this on to see if it would help me see where the problem is.
So at this point I am getting a few more messages, which may help pin down
where the trouble is.  The last message I now get at boot time is coming
from line# 150 in hpt366.c

   pci_bus_clock_list: found match: 0x90c9a731

where the hex number is the value of: chipset_table->chipset_settings
which is about to be returned from the subroutine: pci_bus_clock_list()
This subroutine is only called from 3 places, all of which are inside the
same switch statement. They all head for line 192 where there are only
three statements before the next printk(), which of course never comes
out.  So, assuming that there in nothing sitting in a buffer someplace,
(and I admit that I do not know enough about the console i/o at this stage
of things to know the affects of buffering), then the system is locking up
on one of those 3 statements, which are:

reg2 &= ~0x8000;
pci_write_config_dword(HWIF(drive)->pci_dev, regtime, reg2);
err = ide_config_drive_speed(drive, speed); 

It looks safe to rule out the 1st one. Anyone have any suggestions as to
what is going wrong, or how to fix it?

Best Regards,
Robert Laughlin

HPT366 prevents boot on Abit BE-6 (probably not raid related)

[Robert]

I am still trying to get my HPT366 channels to work on my ABIT BE6
motherboard.  I really appreciate the emails I have gotten to date, but
there is still no joy.  Following in the footsteps of Tim Moore, I
upgraded to 2.2.13 with patches raid0145-19990824-2.2.11, and
ide.2.2.13.1999.patch applied.  This has made no difference.  I
noticed there is a debug option in hpt366.c, namely 
HPT366_DEBUG_DRIVE_INFO
So I turned this on to see if it would help me see where the problem is.
So at this point I am getting a few more messages, which may help pin down
where the trouble is.  The last message I now get at boot time is coming
from line# 150 in hpt366.c

   pci_bus_clock_list: found match: 0x90c9a731

where the hex number is the value of: chipset_table->chipset_settings
which is about to be returned from the subroutine: pci_bus_clock_list()
This subroutine is only called from 3 places, all of which are inside the
same switch statement. They all head for line 192 where there are only
three statements before the next printk(), which of course never comes
out.  So, assuming that there in nothing sitting in a buffer someplace,
(and I admit that I do not know enough about the console i/o at this stage
of things to know the affects of buffering), then the system is locking up
on one of those 3 statements, which are:

reg2 &= ~0x8000;
pci_write_config_dword(HWIF(drive)->pci_dev, regtime, reg2);
err = ide_config_drive_speed(drive, speed); 

It looks safe to rule out the 1st one. Anyone have any suggestions as to
what is going wrong, or how to fix it?

Best Regards,
Robert Laughlin

Re: HELP, switched disks !!

[Robert Dahlem]

Chris,

On Fri, 26 Nov 1999 21:40:37 +0100, KS wrote:

>We just had some memory added in our server that is on some remote location.
>The guy also switched the disks that were in the raid1.
>He connected them to wrong cables, when he realised that sth is wrong,
>he connected them like before and as he say fsck started,
>
>it failed and asked for root password. 
>What should be done in such a situation ?

Well, type in the root password.

Then cat /proc/mdstat and post what's in there.

Also post /etc/raidtab and 'fdisk -l' for your disks.

Regards,
Robert

Re: Trouble autodetecting at boot

[Robert Dahlem]

David,

On Fri, 26 Nov 1999 03:03:35 -0800, David Cunningham wrote:

>Here's the situation.  I've read the Howto.  I've got 2 EIDE disks running
>RAID-1 under Redhat 6.0.  The partition types are fd, the kernel option for
>autodetecting RAID is set, and md0 was built with persistent superblock.
>Here is an excerpt from dmesg:
>
>autodetecting RAID arrays
>(read) hdb1's sb offset: 10432 [events: 0002]
>(read) hdc1's sb offset: 10432 [events: 0002]
>autorun ...
>considering hdc1 ...
>  adding hdc1 ...
>  adding hdb1 ...
>created md0
>bind
>bind
>running: 
>now!
>hdc1's event counter: 0002
>hdb1's event counter: 0002
>request_module[md-personality-3]: Root fs not mounted

Well, you built your kernel with some modules, especially in the md range. Where do 
you 
expect your kernel to load those modules from while not having anything mounted? :-)

Rebuild your kernel with the md parts resident and it won't fail anymore.

Regards,
Robert

Re: Help Raid for sparc

[Robert Dahlem]

James,

On Fri, 26 Nov 1999 08:26:19 -0800 (PST), Mr. James W. Laferriere wrote:

>   I vote for maintaining the Doc's ,  [...]

And do you volunteer to? :-)

Regards,
    Robert

Re: File sizes > 2Gb

[Robert Collier]

In article <01db01bf32eb$5850b1d0$[EMAIL PROTECTED]> you write:
>Do we HAVE to use EXT2 ?  I understand this is a limitation imposed by the
>ext2 filesystem rather than the hardware architecture... Is there not a way
>to create a different file-system using Linux?

EXT2 supports files up to a couple of terrabytes in size, and 64bit linux
can read/write them. However, limitations in the VFS on 32bit linux restrict
32bit linux to 2Gb files.

This limit has been removed in the very latest 2.3 kernels, but we still need
to wait for userspace to catch up.

If you don't need mmap, there is a spacial filesystem called smugfs which allows
large files on 32bit, I don't have URL handy however.

    - Regards, Robert.

-- 
Robert Collier  Windows 2001: "I'm Sorry
[EMAIL PROTECTED]   Dave, I can't do that."

Re: New user troubles: (1) booting (2) size?

[Robert Dahlem]

Stephen,

On Tue, 16 Nov 1999 16:05:19 -0800 (PST), Stephen Walton wrote:

>Well I'm a new convert to software RAID on Linux.  The system I'm
>presently using is a P-150 with Adaptec 2940 SCSI controller, single SCSI
>hard disk which is the boot drive, and a new pair of Western Digital
>Caviar 20.5 GB IDE drives (hda and hdc) acting as software RAID.  This is
>all on RedHat 6.0.
>
>I created my RAID system OK, after setting the system BIOS to put both
>disks in 'normal' mode, which caused both disks to report 16-63-39761 as
>the number of heads-sectors-cylinders.  As a comment:  the docs should
>probably emphasize the importance of shutting down the RAID by hand if
>your system doesn't do it automatically, and RH 6.0 doesn't.
>
>My problems:  first of all, I'm open to suggestions on how to boot this
>system now.  I tried a suggest LILO setup:
>
>boot=/dev/hda
>map=/boot/map
>install=/boot/boot.b
>prompt
>timeout=50
>image=/boot/vmlinuz-2.2.5-22
>   label=linux
>   root=/dev/sda7
>   initrd=/boot/initrd-2.2.5-22.img
>   read-only
>
>to write the boot record to /dev/hda, but LILO still stops after printing
>LI.  The RedHat boot floppy I created will still boot the system and now
>I'm just leaving it in the drive.

Your SCSI disk is not logged in by the BIOS (it will only log in C: AKA 0x80 and D: 
AKA 
0x81). Lilo has no way of determining who the f*** /dev/sda7 is while under control of 
the BIOS, so it will simply fail. While running two IDE disks you will not be able to 
catch and start a kernel from an SCSI disk (its simply not known to the BIOS). Try to 
disable both IDE disks in the BIOS settings and allow your Adaptec to boot from the 
first 
SCSI disk. Under these circumstances you will have to modify your lilo.conf like this:

boot=/dev/sda
disk=/dev/sda bios=0x80
map=/boot/map
install=/boot/boot.b
prompt
timeout=50
image=/boot/vmlinuz-2.2.5-22
label=linux
root=/dev/sda7
initrd=/boot/initrd-2.2.5-22.img
read-only

>The second problem is more serious:  the RAID data are slightly corrupt
>(no proper shutdown before a system reboot), 

As already suggested: Use the new RAID style (partition type 0xfd) and the kernel will 
automatically shutdown the RAID devices.

Regards,
Robert

Re: booting from second drive.

[Robert Dahlem]

Luis,

On Wed, 17 Nov 1999 11:00:43 -0500 (EST), Luis Costabile wrote:

>Thanks Robert.   I'm now able to boot from the boot partition of the
>second drive.

Well, that's what we wanted. :-)

>One question, where do you find out that sdb is supposed to be bios=0x80
>  what about sda ?  

sdb becomes sda in the backup case. You can see lilo reading things like 0x80 and 0x81 
when starting it with '-v -v -v' (very very verbose :-). Its some kind of a well known 
fact that the first recognized disk is adressed by the BIOS as 0x80, the second as 
0x81 
and so on.

What lilo.conf.backup does is to fake that 0x81 (sdb while operating normal) is 
already 
0x80 (while it is not) to prepare it for backup operation (when it is logged in as 
0x80 
by the BIOS).

>is there a table somewhere I could see that info for myself?

Try Ralph Brown's Interrupt List.

BTW: You should consider attaching your second drive as sdc (first drive of second 
controller) thus giving you more redundancy.

Regards,
Robert

Re: booting from second drive.

[Robert Dahlem]

Luis,

On Mon, 15 Nov 1999 12:49:07 -0500 (EST), Luis Costabile wrote:

>Hello, I installed root as a mirrored volume.  I have two SCSI disks only
>in the system.  According to the HOWTO it's suggested to have /boot reside
>on a separate partition on each disk.   I installed the kernels and other
>stuff normally on /boot to these partitions then one at a time ran lilo
>against them.  When I run lilo to the /boot partition on the first drive
>everything is OK,  when I do the same to the other drive I get a warning
>that I'm installing lilo to the second disk drive but as far as I can see,
>LILO is installed anyway.
>
>Now, when I try and boot , everything is fine, system comes up etc, but
>when I disable the first drive and try to boot from the second drive I get
>that LILILILILIL  and bunch of other lilo stuff, none of which makes sense
>to me.

Create a second boot partition on the second drive and mount it as /boot2.

Now copy all files from /boot to /boot2.

After this create a second lilo configuration file (i.e. /etc/lilo.conf.backup)

 boot=/dev/sdb
 disk=/dev/sdb bios=0x80
 map=/boot2/map
 install=/boot2/boot.b

 image=/boot2/vmlinuz
 root=/dev/md0
 label=linux

Now run 'lilo -C /etc/lilo.conf.backup'. This one will probably boot as /dev/sda from 
the 
former /dev/sdb when you remove /dev/sda (indeed I tested this an EIDE configuration 
with 
/dev/hda and /dev/hdc, not with SCSI). Please remark, that it will come up as 
/dev/sda!!!

You will have to make sure that in case of loss of any of your disks your system will 
come up with some reasonable disk names: Lets assume you have /dev/sda, sdb, sdc and 
sdd 
forming a RAID-1 /dev/md0 consisting of /dev/sdb and /dev/sdd. You will be out of the 
game whenever you loose sda or sdb ...

Regards,
Robert

Re: adding/removing linear raid drives

[Robert Dahlem]

Jakob,

On Wed, 10 Nov 1999 19:55:56 +0100, Jakob  stergaard wrote:

>After a change you can mkraid the array again, and if you have an ext2
>filesystem on it, you can probably use ext2resize to resize that to fit
>the new size of the array.   Note however, that ext2resize only supports
>*growing* filesystems.  You're out of luck for shrinking.

That's not _my_ experience. I tried it last weekend with what I found on 
http://www.dsv.nl/
~buytenh/ext2resize (1.0.5 or so) and it happily shrinks ext2 filesystems.

Regards,
Robert

Abort

[Zsiga Robert]

I have a problem.
I would like to create a "raid arrays".I have two SCSI disks.I would like
to use the RAID0 mode.When i mkraid /dev/md0 , mkraid aborted.
Why?
The /etc/raidtab file is the sample file for RAID0.The -f --really-force
don't help.Neither if i recreate the partitions.The second problem the
mkraid write that the partitions consist ext2.
What can i do???
Help me!


      ZSIGA ROBERT  
  

 Lakcim: Berettyoujfalu  E-mail: [EMAIL PROTECTED]
 Zoldfa u. 37.   [EMAIL PROTECTED]
 4100Telefon: 06-54-404-124
 
 Kollegium: Budapest Telefon: 368-20-36 337. szoba
Becsi ut 104-108.
1036

Re: Replacing a RAID-1 mirror drive

[Robert Dahlem]

Matt,

On Wed, 10 Nov 1999 11:18:37 -0500, Matt Kimmel wrote:

>I am running a RAID-1 mirror set of two disks on our Linux server here.  A couple of 
>weeks ago, one of the mirrors went bad.  While waiting for a replacement, I ran the 
>mirror with a single drive, which worked fine.  However, now that I have a 
>replacement 
>drive, I can't seem to add it back into the mirror group.  Here's my /etc/raidtab:

>raiddev /dev/md0
>raid-level  1
>nr-raid-disks   2
>device  /dev/sda1
>raid-disk   0
>device  /dev/sdd1
>raid-disk   1

>It seems like a problem in /dev/sda1's superblock--that it's not indicating /dev/sdd1 
>as 
>another disk in the mirror anymore.  It seems like I should be able to fix this with 
>mkraid, but every mkraid option is described as destructive in the man page.  What I 
>would *like* to do is to add /dev/sdd1 to the RAID-1 set and let it be resynchronized 
>to 
>the contents of /dev/sda1 by the recovery thread.  Is this possible?  Or am I stuck 
>having to recreate the RAID-1 set, make a new file system, and restore all the files 
>by 
>hand?

You have to fdisk /dev/sdd (remember to set the partition type fd on /dev/sdd1), 
reboot 
your system, then do

raidhotadd /dev/md0 /dev/sdd1

Regards,
Robert

RE: [new release] raidreconf utility

[Gerrish, Robert]

> From: Jakob Østergaard [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, November 02, 1999 1:56 PM
> To: Linux RAID mailing list
> Subject: Re: [new release] raidreconf utility
> 
> 
> My only experience with LVM is from HPUX.  I could create the 
> equivalent of RAID-0
> there using LVM only, and it is my understanding that LVM for 
> Linux can do the same.
> It should indeed be possible to create the equivalent of 
> RAID-5 as well, using only
> LVM.  But still the LVM would have to support extending the parity-VG.
> 
> raidreconf will hopefully be useful for people to do these 
> tricks, until the LVM
> gets the needed features (which may be years ahead).
> 

LVM on HPUX handles RAID-1, whereas LVM on linux does not support
mirroring.  HPUX does give you the ability to expand the size of
the RAID-1 devices in a volume group.  

It is my understanding, that you would have to have RAID-5 or RAID-0
under RAID-1 to get this to work in software raid.  Both schemes
are lacking the ability to handle expansion of RAID-1 under linux.

> IMHO the support for redundancy should be in the LVM layer. 
> This would eliminate
> the need for RAID support as we know it today, because LVM 
> could provide the same
> functionality, only even more flexible.  But it will take time.

In the little bit I have been following the linux-lvm mailing
list, it does not look like they are prusuing RAID-1 implementation
with LVM.

Bob Gerrish

RE: ide and hot swap

[Gerrish, Robert]

> Seth Vidal wrote:
>
> We've got DLT's doing backups right now and we're conceiving 
> that it might
> be cheaper to setup a system with 2 or 3 linear striped or 
> raid 0 34+gig
> ide disks and have 2 sets of these disks that we swap out 
> week to week for
> backups - rather than spend a fortune in DLT tapes and deal with a
> whopping 4MB/s transfer time. We would be using a set of disks for 4
> weeks then swapping out to another set - the other set would be fresh
> formatted at that point and would be ready to go for the next month's
> backups.

Price wise, this seems like a good approach.  If it were my system, I would
be concerned about disaster recovery.  I have been a believer for a long 
time in tape rotation and offsite storage.  Also, you are risking losing
4 weeks worth of data; a full backup at least weekly and incremental
backups can save your business.  You not only should think about system
failures,
but fires, floods, etc.  An onsite disk storage scheme doesn't take these
situations into account.  Perhaps, if you want to consider alternate
storage,
you should look at optical media or some other approach.

Bob Gerrish

RE: superblock Q/clarification

[Gerrish, Robert]

> David Cooley wrote:
>
> When I first set up my Raid 5, I made the partitions with Fdisk, and 
> started /dev/hdc1 at block 0, the end was the end of the disk (single 
> partition per drive except /dev/hdc5 is type whole disk).
> It ran fine until I rebooted, when it came up and said there 
> was no valid superblock.  I re-fdisked the drives and re-ran mkraid and 
> all was well until I rebooted.  I read somewhere (can't remember where 
> though) that block 0 had to be isolated as the superblock was written
there... I 
> re-fdisked all my drives so partition /dev/hdx1 started at block 1 instead

> of zero and haven't had a problem since.  I'm running Kernel 2.2.12 with 
>raidtools 0.90 and all the patches..
 
I had a similar experience in setting up Raid 1 devices.  Same Kernel and
patches.  Now my raid disks have nothing in block 0.  For a while, I was
using
mismatched raid kernel and raidtools patches, so that might have been the
point where I tried this solution.  

Bob Gerrish

Re: IP Networking Consultant Wanted for stack port

[Robert]

On Tue, 2 Nov 1999 [EMAIL PROTECTED] wrote:
>   Certainly, but as you have already admitted, your IP stack will be
> a derivative of the Linux IP stack (which is GPL'd), and therefore any
> derivative of a GPL'd package must release its source.

I have no objections to releasing the derived IP stack code.
Robert

IP Networking Consultant Wanted for stack port

[Robert]

This is not really the correct list for this post, but I feel there are
some very knowledgable people on this list, I figure that someone here may
very well be qualified or know of someone else who is..Feel free to
forward this to someone else if the latter is the case.

I am looking for a consultant with Linux Networking experience,
specifically ip experience at the driver level.  The task is to port the
linux ip stack to a z80 based machine.  The z80 machine is being
programmed in C using a cross compiler called UNIWARE.  The z80 operating
system is proprietary, and was developed in house by my company.  We have,
therefore, complete rights and source to it.

To apply, please send an email write-up (in plain text, no HTML or
attachments) with your relevant experience, outline of the scope of effort
you feel is involved, estimate of the cost, and any questions you may
have, directly to my email address.

Best Regards,
Robert Laughlin

Re: Web site on IDE patch status

[Robert]

Andre, Thanks so much for the reply.  In my case I am using a BE6
motherboard made my ABIT.  This mother board has all 4 IDE channels build
into it, so there is no PCI card.  However when looking at the way the
BIOS acts, I think the motherboard emulates a PCI card with this chipset.
So even though there is no plugin card, everything may look as if there is
from a software point of view.  At the moment here is my current setup.
which works at UDMA33:

IDE1Hard Drive(pri)CD rom(sec)
IDE2Hard Drive(pri)Hard Drive(sec)
IDE3 (unattached)
IDE4 (unattached)

Now if I try hooking up an additional hard drive to either IDE3 or IDE4 as
follows, the system hangs:
  
IDE1Hard Drive(pri)CD rom(sec)
IDE2Hard Drive(pri)Hard Drive(sec)
IDE3Hard Drive(pri)
IDE4 (unattached)

At the moment, I am not trying to boot off the IDE3 or IDE4 drive. Below
is the point at which the boot process stops:

HPT366: onboard version of chipset, pin1=1 pin2=2
HPT366: IDE controller on PCI bus 00 dev 98
HPT366: not 100% native mode: will probe irqs later
ide2: BM-DMA at 0xe000-0xe007, BIOS settings: hde:DMA, hdf:pio  
HPT366: IDE controller on PCI bus 00 dev 99
HPT366: not 100% native mode: will probe irqs later
ide3: BM-DMA at 0xec00-0xec07, BIOS settings: hdg:pio, hdh:pio
hda: QUANTUM FIREBALLP KA18.2, ATA DISK drive
hdb: CD-532E-B, ATAPI CDROM drive
hdc: QUANTUM FIREBALLP KA18.2, ATA DISK drive
hdd: QUANTUM FIREBALLP KA18.2, ATA DISK drive
hde: QUANTUM FIREBALLP KA18.2, ATA DISK drive   
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
ide2 at 0xd800-0xd807,0xdc02 on irq 11  
hda: QUANTUM FIREBALLP KA18.2, 17624MB w/371kB Cache, CHS=2246/255/63, UDMA(33)
hdc: QUANTUM FIREBALLP KA18.2, 17624MB w/371kB Cache, CHS=35808/16/63, UDMA(33)
hdd: QUANTUM FIREBALLP KA18.2, 17624MB w/371kB Cache, CHS=35808/16/63, UDMA(33)
(hang)


Please let me know what the next step is.

Best Regards,
Robert Laughlin



 On Sun, 31 Oct 1999, Andre Hedrick
wrote:

> 
> In order to get the off board cards to boot you have to enable the option
> first, now comes the FUN!
> 
> You are now required to determine the PCI card slot order.
> Since LILO does not like to boot drives beyond the fourth device, in IDE
> land.  You must call "pci=reverse" to invert the pci-device scan order.
> 
> For starters, edit you lilo.conf file to default to prompt booting.
> 'linux pci=reverse' will invoke the process.
> 
> I expect the system to hang and it better hang if you have more that just
> the onboard chipset.  If it does not, I need to know.  Will do a follow up
> later once you hang your machine.
> 
> Do not worry, it will never get to a partition check or mount.
> 
> Andre Hedrick
> The Linux IDE guy

Web site on IDE patch status

[Robert]

Andre, as you may remember from a prior email, I am anxious to get my
HPT366 controller chipset working.  At present I am using 2.2.12 with
patches: ide.2.2.12.19991014.patch & raid0145-19990824-2.2.11. The system
fails to boot when a drive is connected to either IDE3 or IDE4. In my
case, it's more important to me that I can use the 3rd and 4th channels,
than whether they run at 33 or 66. And of course I would be happy to
supply additional information to you at any time.  But I know you must be
a busy guy.

Anyway here is my question:

Do you have a web site where I check in from time to time, to see the
current status of your efforts and the latest patches and what they
address:  Here is the web address I am using at the moment, is there a
better one?   ftp://ftp.us.kernel.org/pub/linux/kernel/people/hedrick/

Best Regards,
Robert Laughlin


On Sat, 30 Oct 1999, Andre Hedrick wrote:
> Hacking is easyDoing it correct is pain.

RE: Compaq Smart Array & RH6.1

[Gerrish, Robert]

>Hardware RAID help, please. Thanks.
>  
>I have a Compaq ProLiant 2500 Server using 2x PPro200mhz, 128MB RAM, 
>Compaq SmartArray 2/E EISA Card with 2x 4.3G and 1x 2.0G HotSwappable HD.
>but I cannot install RedHat 6.1 or SuSE 6.2, they both cannot see my Smart
2/E.
>Under RedHat 6.1, I tried to use "Expert" installation mode, and select
SCSI, 
>then select "Compaq Array", give it parameter 0xC8000, machine just stay
there...
>same under SuSE 6.2, choose Kernel Driver, and load Compaq Array, giving
0xC8000, 
>just stay there... or else, if I don't mention "Compaq Array", they both
cannot 
>see any of my HD.

I have been experiencing similar problems.  I tried using the kernel &
initrd.img I am
using to boot the machine (kernel 2.2.12 with latest raid patches) but the
boot fails
with a kernel panic when I superimpose these on the RedHat 6.1 bootnet
image.

>Anyone experience with this Compaq SmartArray? 
> 
>I read the 2.2.12 kernel source document, it said I need to recompile the
kernel 
>with "patch"?! does anyone know do I need that patch? but how can I patch
it if 
>I cannot install it?
>  
To get 6.0 going, which was supposed to support the Smart2, I had to compile
a 
custom kernel with the compaq patches (smart2-1.0.5.patch &
sim710-990829_diff) 
applied.  

I would try the 2.2.12 kernel with the latest those paches and the raid
patches applied.
Get it to boot the image on a floppy, transfer that kernel & initrd to a
copy of the RedHat
floppy and it should work.

I have to get other machines in production for Y2K relief (retiring novel &
going to linux / NT)
this weekend, so I will not get to try it this week.  If you have success
let me (and the list)
know what you did.

>I had good experience with Software-RAID on both different machine. 
>this is the first machine that I had a hardware-RAID controller.

Compaq is difinetely different and NOT fun :-(
>  
Bob

RE: RedHat 6.1 Upgrade

[Gerrish, Robert]

>I haven't done the upgrade myself but I have seen 2 things regarding this:
>
>1. RH has already released 2 new images (boot & updates) that are supposed
>to fix some bugs w/ the original install

I downloaded their current release on Tuesday.  It appears that they still
have 
a few bugs with upgrade.  I realize that old style raid would be a problem,
but
I am on the current release patches to kernel 2.2.12 and raidtools 0.90.

I need the machine running in production by tonight, so I may have to let
this one lay.  I am partially on RH 6.1 and partially on Mandrake 6.0.

>2. The RAID stuff is only supported in the GUI install, not text.

I had a partial(?) GUI install.  I started with their boot image in images.
Thanks you all for the help so far.

>> -Original Message-
>> From:Gerrish, Robert [SMTP:[EMAIL PROTECTED]]
> >> Sent:  Thursday, October 28, 1999 4:35 PM
>> To:  [EMAIL PROTECTED]
>> Subject: RedHat 6.1 Upgrade
>> 
>> I am trying to install RedHat 6.1 as an upgrade.  The install recofnizes
>> my RAID1 drives as "Linux RAID", but when I do an upgrade, it tells
>> me that I have no linux partitions and it can't proceed.  Has anyone
>> got a work around for this.  The only solution I could see was perhaps
>> installing on the raw hd partitions and then relinking the mirrors.

> Bob Gerrish  -  [EMAIL PROTECTED]

RedHat 6.1 Upgrade

[Gerrish, Robert]

I am trying to install RedHat 6.1 as an upgrade.  The install recofnizes
my RAID1 drives as "Linux RAID", but when I do an upgrade, it tells
me that I have no linux partitions and it can't proceed.  Has anyone
got a work around for this.  The only solution I could see was perhaps
installing on the raw hd partitions and then relinking the mirrors.

Bob Gerrish  -  [EMAIL PROTECTED]

Re: moving /lib - off topic advice wanted

[Robert]

Nope, sure don't.  My new /lib *partition* mount is not even in
/etc/fstab. What will happen at boot time is the /lib *directory* on the /
partition will be available.  Right? I never deleted the data in that
partition. I just mounted my new /lib *partition* on top of it for test.

Best Regards,
Robert Laughlin


On Wed, 20 Oct 1999, Seth Vidal wrote:

> > Perhaps I am wrong, I expected that a reboot would make the original /lib
> > available again at boot time.  The data is still there, just hidden by the
> > mount, right?  
> 
> mounting only occurs after fstab is processed.
> 
> you can't process fstab with the mount command if there are no libraries
> for mount to rely on.
> 
> see the problem?
> 
> -sv
> 
> 
> 
> 

Re: moving /lib - off topic advice wanted

[Robert]

On Wed, 20 Oct 1999, Seth Vidal wrote:
> I think unless your drive mounting and init binaries are statically
> linked you're going to hit trouble at boot time.

Perhaps I am wrong, I expected that a reboot would make the original /lib
available again at boot time.  The data is still there, just hidden by the
mount, right?  

Best Regards,
Robert Laughlin

moving /lib - off topic advice wanted

[Robert]

I suppose this is an off topic question, other than it's related to
disks...:)   I find that my / partition is more full than I would like.  I
like to keep extra space available in case something unexpected happens.
Here is the current setup:

Filesystem   1k-blocks  Used Available Use% Mounted on
/dev/hda862187 48208 10768  82% /
/dev/hda1 7746  3955  3391  54% /boot
/dev/hdc5   495714 61545408568  13% /home
/dev/hda10 1274632690633518135  57% /usr
/dev/hda746632 18515 25709  42% /var   

Turns out /lib is taking about 30 megs, so my first idea is to make a new
partition for /lib.  I have done this:

Filesystem   1k-blocks  Used Available Use% Mounted on
/dev/hda862187 48208 10768  82% /
/dev/hda1 7746  3955  3391  54% /boot
/dev/hdc5   495714 61545408568  13% /home
/dev/hda10 1274632690633518135  57% /usr
/dev/hda746632 18515 25709  42% /var
/dev/hda6   116630 30565 80043  28% /lib  

I figured this was a good first step, because if anything went wrong, I
could just reboot and the old /lib directory would still be there:)
(I have not put the /lib mount into /etc/fstab yet.)

So here are my questions, is it safe for me to do the following steps:

1. reboot to get /dev/hda6 unmounted (it's busy now, so umount does not
work)

2. mv /lib /lib.old
3. mkdir /lib
4. chmod 755 /lib
5. chown 0.0 /lib
6. mount /dev/hda6 /lib 
7. rm -rf /lib.old  {after testing}

Or am I going to run into trouble because /lib's files will be unavailable
for a bit while I enter these commands?  Is there a better way to enlarge
/?  In general how to you recommend changing partition sizes?  Is this an
argument for not seperating directories into different partitions, since
it's harder to keep the free space evenly distributed?

Best Regards,
Robert Laughlin

RE: how to re-introduce a spare?

[Gerrish, Robert]

> Michael Franzino wrote:
> 
>I'm new to Linux and this discussion group so
>please be patient.

I'm fairly new to this group, also, but I have had 
considerable experience with RAID 1 on Sun OS 
& Solaris.

> . . . It is the three 
>SCSI drives that make up the mirrored pair 
>plus one spare. I want to use cron to bring-down
>one of the drives in the mirrored pair every hour.
>The reconstruction will then bring in the spare as 
>the new half of the mirrored pair. The idea (likely
>bad) is that this makes sure I always have a good 
>filesystem no more than one hour old on the drive that
>has been brought down? To make this work I need help 
>with following questions:
>  
If it were my system, I would prefer to have a three way mirror
with at least two (ideally, three) separate controllers.
I would want to detach a mirror to backup the system to tape, 
daily,  giving me a snapshot of the system.  If I were really 
paranoid, I would detach and reattach this mirror several times 
a day for backup.  This would make a less vulnerable 
system and the type of data security you are looking for.

Reconstructing as you suggest leaves you with one drive on line, 
one rebuilding and one with data that is growing stale as you reconstruct 
the new one.  The whole point of RAID 1 is to have redundant data
storage.

If you detached to backup the system, you would still have your
questions:

>1) Once a spare has become part of the mirrored
>   pair (after reconstruction), how do I re-introduce
>   the third drive back as the new spare? Are there 
>   console commands for this?

This would change to how do I detach and reattach a sub mirror.

Bob Gerrish

Problems with HPT366

[Robert]

I am having trouble booting while using IDE channels 3 & 4. My motherboard
is a BE6, with a celeron processer. I am running 2.2.12 straight from the
tarball, with two patches applied:

1. raid0145-19990824-2.2.11  to add the lastest raid patches
2. ide.2.2.12.19991014.patch to add lastest IDE patches

(BTW, I was not able to apply these patches in the opposite order. The
kernel would not compile after doing so.)  I have attempted to email Andre
Hedrick, who is apparnetly the author of the IDE patch I used, but the
address in the CREDITS file: [EMAIL PROTECTED] comes back
with "user unknown".  

Does anyone know how to get in touch with him, or an alternate contact
point for this problem?

Best Regards,
Robert Laughlin

Re: raidtools kernel patch

[Robert]

Well, That's interesting, if I apply the patches in the opposite order,
they work.  Perhaps we need a road map as to what order patches need to be
applied in?

Best Reagrds,
Robert


On Sat, 16 Oct 1999, Robert wrote:

> Hm, those 2.2.11 raid patches did not work for me.  After applying and
> ignoring the 2 errors, I got this upon compiling:
> 
> gcc -D__KERNEL__ -I/usr/src/linux-2.2.12/include -Wall -Wstrict-prototypes
> -O2 -fomit-frame-pointer -fno-strict-aliasing -pipe -fno-strength-reduce -m486
> -malign-loops=2 -malign-jumps=2 -malign-functions=2 -DCPU=586   -DEXPORT_SYMTAB
> -c md.c
> md.c:96: `DEV_MD_SPEED_LIMIT' undeclared here (not in a function)
> md.c:96: initializer element for `md_table[0].ctl_name' is not constant
> md.c:102: `DEV_MD' undeclared here (not in a function)
> md.c:102: initializer element for `md_dir_table[0].ctl_name' is not constant
> make[3]: *** [md.o] Error 1
> make[3]: Leaving directory `/usr/src/linux-2.2.12/drivers/block'
> make[2]: *** [first_rule] Error 2
> make[2]: Leaving directory `/usr/src/linux-2.2.12/drivers/block'
> make[1]: *** [_subdir_block] Error 2
> make[1]: Leaving directory `/usr/src/linux-2.2.12/drivers'
> make: *** [_dir_drivers] Error 2 
> 
> 
> This was a stock 2.2.12 kernal from the tar-ball, with only one other
> patch applied first: ide.2.2.12.19991014.patch
> After applying the IDE patch, the kernel did compile without error.  But
> after then applying the patch raid0145-19990824-2.2.11 the compile failed
> as shown above.  Anyone have any thoughts, suggestions, ideas ?
> 
> Thanks in advance,
> Robert
> 
> 
> 
> 
> On Wed, 13 Oct 1999, Marc Haber wrote:
> 
> > In ka.lists.linux.raid, you wrote:
> > >I run a 2.2.12 kernel. Is the 2.2.11 raid patch included in that or do I
> > >need to use the 2.2.11 patch on the 2.2.12 source tree? Or should I
> > >downgrade my kernel? Also, as to the question I previously posted, I am
> > >using the most current set of raid tools.
> > 
> > You can use the 2.2.11 patch on 2.2.12. However, the patch will
> > complain but the changes the patch wants to do are already there. You
> > can safely ignore these messages.
> > 
> > Greetings
> > Marc
> > 
> > -- 
> > -- !! No courtesy copies, please !! -
> > Marc Haber  |   " Questions are the | Mailadresse im Header
> > Karlsruhe, Germany  | Beginning of Wisdom " | Fon: *49 721 966 32 15
> > Nordisch by Nature  | Lt. Worf, TNG "Rightful Heir" | Fax: *49 721 966 31 29
> > 
> 

Re: raidtools kernel patch

[Robert]

Hm, those 2.2.11 raid patches did not work for me.  After applying and
ignoring the 2 errors, I got this upon compiling:

gcc -D__KERNEL__ -I/usr/src/linux-2.2.12/include -Wall -Wstrict-prototypes
-O2 -fomit-frame-pointer -fno-strict-aliasing -pipe -fno-strength-reduce -m486
-malign-loops=2 -malign-jumps=2 -malign-functions=2 -DCPU=586   -DEXPORT_SYMTAB
-c md.c
md.c:96: `DEV_MD_SPEED_LIMIT' undeclared here (not in a function)
md.c:96: initializer element for `md_table[0].ctl_name' is not constant
md.c:102: `DEV_MD' undeclared here (not in a function)
md.c:102: initializer element for `md_dir_table[0].ctl_name' is not constant
make[3]: *** [md.o] Error 1
make[3]: Leaving directory `/usr/src/linux-2.2.12/drivers/block'
make[2]: *** [first_rule] Error 2
make[2]: Leaving directory `/usr/src/linux-2.2.12/drivers/block'
make[1]: *** [_subdir_block] Error 2
make[1]: Leaving directory `/usr/src/linux-2.2.12/drivers'
make: *** [_dir_drivers] Error 2 


This was a stock 2.2.12 kernal from the tar-ball, with only one other
patch applied first: ide.2.2.12.19991014.patch
After applying the IDE patch, the kernel did compile without error.  But
after then applying the patch raid0145-19990824-2.2.11 the compile failed
as shown above.  Anyone have any thoughts, suggestions, ideas ?

Thanks in advance,
Robert




On Wed, 13 Oct 1999, Marc Haber wrote:

> In ka.lists.linux.raid, you wrote:
> >I run a 2.2.12 kernel. Is the 2.2.11 raid patch included in that or do I
> >need to use the 2.2.11 patch on the 2.2.12 source tree? Or should I
> >downgrade my kernel? Also, as to the question I previously posted, I am
> >using the most current set of raid tools.
> 
> You can use the 2.2.11 patch on 2.2.12. However, the patch will
> complain but the changes the patch wants to do are already there. You
> can safely ignore these messages.
> 
> Greetings
> Marc
> 
> -- 
> -- !! No courtesy copies, please !! -
> Marc Haber  |   " Questions are the | Mailadresse im Header
> Karlsruhe, Germany  | Beginning of Wisdom " | Fon: *49 721 966 32 15
> Nordisch by Nature  | Lt. Worf, TNG "Rightful Heir" | Fax: *49 721 966 31 29
> 

Kernel panic in Raid...

[Robert (Drew) Norman]

I got the following error on my redhat 6.1 linux box.  It appears that
something is wrong with the raid code in the kernel?  I recieved a similar
message on a separate machine yesterday running 2.2.10.

Please advise.

--Drew


Oct 13 12:12:30 dhcp-rtp-34-118 kernel: Unable to handle kernel paging
request at virtual address 245c8b53
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: current->tss.cr3 = 11b21000, %cr3
= 11b21000
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: *pde = 
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: Oops: 
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: CPU:0
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: EIP:0010:[find_buffer+104/144]
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: EFLAGS: 00010206
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: eax: 245c8b53   ebx: 0007
ecx: 0002ff84   edx: 245c8b53
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: esi: 000d   edi: 0900
ebp: 823f   esp: d1b23e38
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: ds: 0018   es: 0018   ss: 0018
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: Process make (pid: 751, process
nr: 42, stackpage=d1b23000)
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: Stack: 823f 0900 0002ff84
c0126a87 0900 823f 1000 c0126cee
Oct 13 12:12:30 dhcp-rtp-34-118 kernel:0900 823f 1000
823f d1b4e550  d1b4e550 40015000
Oct 13 12:12:30 dhcp-rtp-34-118 kernel:c013e013 0900 823f
1000 d1b23f4c   d1b4e550
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: Call Trace: [get_hash_table+23/36]
[getblk+30/324] [inode_getblk+71/392] [ext2_getblk+194/524]
[ext2_bread+35/276] [ext2_readdir+153/1012] [handle_mm_fault+192/332]
Oct 13 12:12:30 dhcp-rtp-34-118 kernel:[do_page_fault+290/796]
[sys_getdents+159/228] [filldir+0/132] [system_call+52/56]
Oct 13 12:12:30 dhcp-rtp-34-118 kernel: Code: 8b 00 39 6a 04 75 15 8b 4c
24 20 39 4a 08 75 0c 66 39 7a 0c

autodetect help needed

[Gerrish, Robert]

I need some help getting autodetect working.  I've tried everything and am
probably just missing some little piece.

I am running Mandrake 6.0.  I have a custom kernel, 2.2.12 with the raid0145
patch set and raidtools 0.90 (and lvm support).  Raid is set up in the
kernel 
with the CONFIG_MD_BOOT=y.  I had wanted RAID 1 across controllers. 
I have tried RAID 1 and RAID 0 and RAID 0+1.  None of the configurations
I have tried will autodetect.

I have 3 IDE disks.  /dev/hda is small (400MB) and has a boot sector and 
a 20MB /boot partition.  I also have an 18GB SCSI on an adaptec 2940U2W.  
I am currently using the root partitioin from /dev/sda1.

Raid comes up on any of the configurations I have tried when
I use raidstart.  I can mkfs, mount, etc.  They work well, they
just won't autodetect the drives.

/etc/raidtab:

raiddev /dev/md1
raid-level  0
nr-raid-disks   2
nr-spare-disks  0
persistent-superblock   1
chunk-size  4
device  /dev/hdb2
raid-disk   0
device  /dev/hdb3
raid-disk   1


raiddev /dev/md3
raid-level  0
nr-raid-disks   2
nr-spare-disks  0
persistent-superblock   1
chunk-size  4

device  /dev/hdc2
raid-disk   0

device  /dev/hdc3
raid-disk   1

Originally I tried RAID 1 with:

(original /etc/raidtab was)

raiddev/dev/md0
raid-level 1
nr-raid-disks  2
nr-spare-disks 0
persistent-superblock   1
chunk-size 4
device /dev/hdb2
raid-disk  0
device /dev/hdc2
raid-disk  1

Current partitions on hdb and hdc are:

Disk /dev/hdb: 15 heads, 63 sectors, 8894 cylinders
Units = cylinders of 945 * 512 bytes

   Device BootStart   EndBlocks   Id  System
/dev/hdb1 2   140 65677+  82  Linux swap
/dev/hdb2   141   357102532+  fd  Unknown
/dev/hdb3   358   574102532+  fd  Unknown

Disk /dev/hdc: 15 heads, 63 sectors, 8940 cylinders
Units = cylinders of 945 * 512 bytes

   Device BootStart   EndBlocks   Id  System
/dev/hdc1 2   140 65677+  82  Linux swap
/dev/hdc2   141   357102532+  fd  Unknown
/dev/hdc3   358   574102532+  fd  Unknown

The original configurations for hdb and hdc were as 1 250MB
partition on each drive in partition 1 (hdb1 & hdc1).  I left the
first cylinder for the persistent superblock.  Do I need the first
cylinder to be on a file system?  

cat /proc/mdstat

Personalities : [linear] [raid0] [raid1] [raid5]
read_ahead not set
unused devices: 

/etc/lilo.conf

disk=/dev/hda bios=0x80
boot=/dev/hda
map=/boot/map
install=/boot/boot.b
prompt
timeout=50
image=/boot/vmlinuz-2.2.12-md-lvm
label=linux
root=/dev/sda1
initrd=/boot/initrd-2.2.12-md-lvm
read-only
image=/boot/vmlinuz-2.2.12-md-lvm
label=raid
root=/dev/md0
initrd=/boot/initrd-2.2.12-md-lvm
read-only

(I was trying to use md0 as root device, but I know that autodetect
has to work first.)

dmesg shows:

FDC 0 is a post-1991 82077
md driver 0.90.0 MAX_MD_DEVS=256, MAX_REAL=12
linear personality registered
raid0 personality registered
raid1 personality registered
raid5 personality registered
raid5: measuring checksumming speed
raid5: MMX detected, trying high-speed MMX checksum routines
   pII_mmx   :  1001.649 MB/sec
   p5_mmx:  1045.464 MB/sec
   8regs :   774.573 MB/sec
   32regs:   411.480 MB/sec
using fastest function: p5_mmx (1045.464 MB/sec)
(scsi0)  found at PCI 11/0
(scsi0) Wide Channel, SCSI ID=11, 32/255 SCBs
(scsi0) Downloading sequencer code... 374 instructions downloaded
scsi0 : Adaptec AHA274x/284x/294x (EISA/VLB/PCI-Fast SCSI) 5.1.19/3.2.4
   
scsi : 1 host.
(scsi0:0:0:0) Synchronous at 40.0 Mbyte/sec, offset 31.
  Vendor: QUANTUM   Model: QM318000TD-SW Rev: N1B0
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sda at scsi0, channel 0, id 0, lun 0
scsi : detected 1 SCSI disk total.
SCSI device sda: hdwr sector= 512 bytes. Sectors= 35566499 [17366 MB] [17.4
GB]
3c59x.c:v0.99H 11/17/98 Donald Becker
http://cesdis.gsfc.nasa.gov/linux/drivers/vortex.html
tulip.c:v0.89H 5/23/98 [EMAIL PROTECTED]
eth0: Lite-On 82c168 PNIC at 0xd000, 00 a0 cc 54 bf b4, IRQ 10.
eth0:  MII transceiver found at MDIO address 1, config 3000 status 7829.
eth0:  Advertising 01e1 on PHY 1, previously advertising 01e1.
Partition check:
 sda: sda1 sda2 < sda5 sda6 sda7 sda8 sda9 >
md.c: sizeof(mdp_super_t) = 4096
autodetecting RAID arrays
autorun ...
... autorun DONE.
VFS: Mounted root (ext2 filesystem) readonly.
Freeing unused kernel memory: 36k freed
Adding Swap: 265032k swap-space (priority -1)
hda: ST3491A D, ATA DISK drive
hdb: ST34321A, ATA DISK drive
hdc: FUJITSU MPD3043AT

HPT366 & DMA mode

[Robert]

Two questions for you experts.  I really appreciate the help I got on
getting my RAID devices working.

Situation: My motherboad is a BE6 from Soft Switch.  It features 4 IDE
connectors.  IDE 1 & 2 were detected at install time and are running
with a CDrom drive and 3 HDs.  IDE 3 & 4 use the Ultra ATA/66 IDE chip
set.  I have been unable to get IDE3 or 4 to work.

Question 1:  The drives on IDE 1 & 2 are presently configured (as reported
by hdparm) to not be using DMA mode.  Needless to say the cpu utilization
can get quite high when doing copies from disk to disk, and generally
performance is barely adequate.  The drives themselves are ATA/66
capable. Can I enable the DMA mode on these disks while they are connected
to IDE1 & 2?

Question 2: Can I get the drives to work at all on IDE3 & 4?  I mean that
even if they work at ATA/33 speeds that would still be a huge improvement
over what is happening now?  I notice that Kernal 2.3 has added support
for IDE HPT343.  I am not sure what the relationship is between this and
what I have, but it sounds like my chip set may not be supported.  But
perhaps it is supported, but only in some degraded mode?

Thanks in advance for any suggestions or ideas or pointers,
Robert

Re: raid as modules?

[Robert]

On Mon, 4 Oct 1999, Martin Bene wrote:
> Make very, very sure that you're actually running your new patched kernel;
> just building the raid modules and trying to load them with your old kernel
> will NOT work.
> 
> This is the classic Kernel Raid 0.36, raidtools 0.90 conflict. Here's the
> much more descriptive error message from the nextrelease of mkraid (patch
> by Tom Livingston):
> 
> ERROR: Unsupported md version in kernel!
> mkraid has determined that you are using a kernel that includes an older
> version of the RAID patches than this version of mkraid supports. Your
> kernel RAID version is 0.36, while mkraid requires a version of 0.90
> or higher.  Newer raid kernel patches can be obtained from:
> ftp://ftp.XX.kernel.org/pub/linux/daemons/raid/alpha/
> where XX is a country code, like US

Is there a way to determine what version of RAID is in my kernel?  I have
two RAID partitions, which both appear to be working. As to the tools, I
get this:

/root> mkraid --version
mkraid version 0.90.0  
/root> raidstart --version
raidstart v0.3d compiled for md raidtools-0.90 
/root> raidstop --version
raidstop v0.3d compiled for md raidtools-0.90   

Best Regards,
Robert

RE: undocumented error from /proc/mdstat: read_ahead not set

[Gerrish, Robert]

I was trying that . . . using raidtools 0.90 without patching the kernel.

I finally got smart earlier today and download the kernel patches, and now
it works well.  I was getting 
exactly the same error messages.  A lot of the documentation implies that
the newer kernels
support RAID as is, especially Red Hat, but that seems to be wrong.  The
kernel patches are quite 
necessary.

I started off with the Mandrake 6.0 and the raidtools 0.90 RPM.  It can be
frustrating.  If you are using
something like that, you should probably download the latest stable kernel
since you have to rebuild it.  

If I have time, I may try to put together an RPM for Red Hat / Mandrake 6.0.
They are using 2.2.9 and I
upgraded to 2.2.12, so it could be a major task.

Bob

> -Original Message-
> From: Jones, Clay [SMTP:[EMAIL PROTECTED]]
> Sent: Friday, October 01, 1999 7:52 AM
> To:   '-_ sergio _-'; [EMAIL PROTECTED]
> Subject:  RE: undocumented error from /proc/mdstat: read_ahead not set
> 
> You are using the "new" raid tools (version 90) without having the kernel
> patched with the "new" raid code. 
> 
> Kernel patches are found in 
> 
>  
> 
> Clay 
> 
> -Original Message- 
> From: -_ sergio _- [ ] 
> Sent: Friday, October 01, 1999 6:42 AM 
> To: [EMAIL PROTECTED] 
> Subject: undocumented error from /proc/mdstat: read_ahead not set 
> 
> 
> hello, i am trying to establish a raid0 array on my two ide disks, 
> but: 
> 
> [root@serek ~]# mkraid --really-force /dev/md0 
> DESTROYING the contents of /dev/md0 in 5 seconds, Ctrl-C if unsure! 
> handling MD device /dev/md0 
> analyzing super-block 
> disk 0: /dev/hdc1, 26588488kB, raid superblock at 26588416kB 
> disk 1: /dev/hdd1, 26588488kB, raid superblock at 26588416kB 
> mkraid: aborted, see the syslog and /proc/mdstat for potential clues. 
> 
> [root@serek ~]# cat /proc/mdstat 
> Personalities : [1 linear] [2 raid0] 
> read_ahead not set 
> md0 : inactive 
> md1 : inactive 
> md2 : inactive 
> md3 : inactive 
> 
> what does mean "read_ahead not set"? i am rather sure that everything 
> with my disks is all right. 
> 
> i am not on list, so please answer on priv email. 
> 
> thanks for raidtools! :-) 
> 
> serek 
> -- 
>  | on_irc: serek at #serek | icq#27579212 
> __ 
> I'd rather have two girls at 21 each than one girl at 42. 
> -- W.C. Fields 
> 

Re: RAID-0 or RAID-linear on Redhat 6.0

[Robert]

I got a few hints off the mailing list, but not all the information I
needed, so last night I just decided to try and make it work with a couple
of test partitions. Eventually I did get raid-0 to work, by just trying
various things.  The biggest thing I found was that you (apparnetly) need
to run mkraid on level-0 even though the man pages say you do not need to.

These were the steps I did:

1. used cfdisk to create two new partitions.

2. rebooted the computer (this seemed to be necessary for the OS to learn
about the new partitions)

3. mkraid --force /dev/md0
This appears to be necessary even for raid-0, even though "man mkraid"
says otherwise). I also put the "chunk-size" parameter in the /etc/raidtab
even though the examples in "man raidtab" show it's not needed, it seems
to, in fact, be needed.  Here is my complete /etc/raidtab:

raiddev /dev/md0
raid-level  0
nr-raid-disks   2   
nr-spare-disks  0
chunk-size  4

device  /dev/hdc7
raid-disk   0
device  /dev/hdc8
raid-disk   1

4. raidstart /dev/md0   
This gives an error message that "dev/md0" already exists.  Apparently
this is ok that it already exists.
 
5. mke2fs /dev/md0
6. mount /dev/md0 /mnt/raid

If anyone sees somthing wrong or not needed with the above, I would
appreciate hearing about it.


This morning an email posted a URL with some very good info:

http://www.ostenfeld.dk/~jakob/Software-RAID.HOWTO

That URL cleared up a lot of stuff for me.  So now I am trying to figure
out how to move my files into a RAID partition.  See, I have about 10 gigs
of data that I want to put into a raid-0 and the disk that has the 10 gigs
on it needs to be part of that raid-0.

Best Regards,
Robert



On Thu, 30 Sep 1999, Paul VanDyke wrote:

> I got my stuff to work... I had to change some things in my /etc/raidtab
> file...
> 
> What is your situation?? Give details on your /etc/raidtab file and disk
> layouts..
> 
> Paul
> 

RAID-0 or RAID-linear on Redhat 6.0

[Robert]

I can't seem to get RAID-0 (or RAID-linear) to work with Redhat 6.0.  It
looks like the tools have been renamed (raidadd is obs, and raidstart is
used instead, the setups are put into /etc/raidtab.)  I was able to get
RAID-1 and RAID-5 to work, but not RAID-0, and I can't seem to find even a
mention about RAID-linear.  If anyone has been successful in getting these
to work, would you please send the configs and setup instructions.  Sorry
if this question has already been answered, I just found and joined this
list.  My latest attempt with /etc/raidtab is shown below.

Thanks in advance for any assistance,
Robert


raiddev /dev/md0
raid-level  0
nr-raid-disks   2   
nr-spare-disks  0
chunk-size  4096

device  /dev/hda5
raid-disk   0
device  /dev/hdc1
raid-disk   1

Patch and raid tools for 2.2.12

[Robert E. Lee]

The latest kernel patch and raid tool combination I found on
ftp.us.kernel.org/pub/linux/daemons/raid/alpha was for the 2.2.11 kernel.
Where might I find the 2.2.12 kernel patches?

Thanks,

Robert E. Lee
Sun Enterprise Certified Engineer
Network Engineer, Access GEITD
Work Phone: (303) 544-6519

mkraid fails while creating a new raid 0 array

[Robert M. Albrecht]

Hi,

I'm using RedHat 6.0 on a Pentium II system. It has three scsi-harddisks. I
use the first one to boot the systems and want to create a raid on the two
other disks.

I've created an /etc/raidtab:

raiddev /dev/md0
raid-level  0
nr-raid-disks   2
persistent-superblock   1
chunk-size  16
device  /dev/sdb1
raid-disk   0
device  /dev/sdd1
raid-disk   1

My /proc/mdstat:

Personalities : [2 raid0]
read_ahead not set
md0 : inactive
md1 : inactive
md2 : inactive
md3 : inactive

When running mkraid --bogus /dev/md0 this happens:

handling MD device /dev/md0
analyzing super-block
disk 0: /dev/sdb1, 102384kB, raid superblock at 102272kB
disk 1: /dev/sdd1, 102384kB, raid superblock at 102272kB
mkraid: aborted, see the syslog and /proc/mdstat for potentional clues

There are no messages in syslog. I'm not sure what the "incative" in mdstats
means.

What's going wrong ? I've tried several kernels (2.2.5 2.2.9 2.2.12
2.2.13pre10 and 2.3.18x). I've played around with some parameters like
chunksize and persistantsuperblock. I'm using raidtools-0.90-5.rpm. I've
also compiled the raidtools myself, but nothing changed.

What to do ? Are there any verbose-switches to get some more messages ?

cu romal

Yet Another RAID0 question...

[Robert McPeak]

Hi - 

 
I'm having 
difficulty getting software RAID0 to work on the following 
configuration:
/dev/hda    - / partition, contains all 
boot, system files.
/dev/hdb    - disk 1 of 
RAID0
/dev/hdc- hardlinked to 
/dev/cdrom
/dev/hdd    - disk 2 of 
RAID0
 
Both drives are 
Western Digital 18gb Ultra/66 drives (AC418000).  Kernel version is 
2.2.9-19mdk (Mandrake 6.0 recompiled with RAID0 support built in, not 
modular).  
 
My /etc/raidtab file 
looks like this:


raiddev 
/dev/md0raid-level  
0#persistent-superblock   
1chunk-size  
32#nr-raid-disks   
2nr-spare-disks  
0#device  
/dev/hdb1raid-disk   
0#device  
/dev/hdd1raid-disk   
1
 
mkraid 0.90.0 returns the 
following:

handling MD device 
/dev/md0analyzing super-blockdisk 0: /dev/hdb1, 17619493kB, raid 
superblock at 17619392kBdisk 1: /dev/hdd1, 17619808kB, raid superblock at 
17619712kBmkraid: aborted, see the syslog and /proc/mdstat for potential 
clues.
 

/proc/mdstat looks like 
this:

Personalities 
: [2 raid0]read_ahead not setmd0 : inactivemd1 : inactivemd2 : 
inactivemd3 : inactive
 
syslog doesn't 
contain any information that I recognize as pertinent, with the exception of the 
message that indicated that the RAID0 personality was loaded at 
boot...
 
 
I've looked at all 
the docs on this that I can find, which appear to be out of date, and refer 
to tools that don't exist in the newer raidtools packages.  I've created 
the /dev/hdb1 and /dev/hdd1 partitions both as type 83 (ext2) and fd (so 
that they automount), but I can't get the RAID to start with either type, and 
it's driving me nuts.  Are the 18gb drives too big for software RAID?  
Can anyone give me a clue as to what the heck is wrong here?  
Thanks!
 
 
Rob

No Subject

[Robert Purdy]

Does anyone know how far ext3 is off? Also I heard SGI had 
released thiercode for journaling? Is this to be adopted into 
ext3?CheersRob

Re: seeking advice for linux raid config

[Robert Collier]

In article <[EMAIL PROTECTED]> you write:
>there's another problem with linux that hasn't been mentioned at the list:
>with ext2 the maximum filesize is 2 GByte. Maybe there's a patch out there
>but i don't think so because it is a design limitation.

Thats not quite correct. The maximum filesize is 2 GByte on a 32bit system.
If you have an alpha or (possibly sparc64?) then you can have much larger
files. This is due to 32bit limits in the vfs (IIRC).

It's also planned to be fixed in 2.3 to allow large file access on 32bit
machines.

- Rob.

-- 
Robert Collier  Smile, it makes people wonder
[EMAIL PROTECTED]   what you are thinking...

problems with raidtools 0.90

[Robert Jones]

I am having trouble bringing up a raid 0 volume.  I followed the instructions in Jakob 
OEstergaard's document and read the man pages, but I am still getting problems.  When 
I run the mkraid I get the following output:

[root@ahab /etc]# /sbin/mkraid --really-force /dev/md0
DESTROYING the contents of /dev/md0 in 5 seconds, Ctrl-C if unsure!
handling MD device /dev/md0
analyzing super-block
disk 0: /dev/sdc1, 8883913kB, raid superblock at 8883840kB
disk 1: /dev/sdd1, 8883913kB, raid superblock at 8883840kB
disk 2: /dev/sde1, 8883913kB, raid superblock at 8883840kB
disk 3: /dev/sdf1, 8883913kB, raid superblock at 8883840kB
disk 4: /dev/sdb2, 8883913kB, raid superblock at 8883840kB
mkraid: aborted

Note that I had to use the --really-force option because I used the older raidtools 
software to see if they worked.  They did, but I didn't know how to take the 
filesystem info off so I have to use this option.

Here is my /etc/raidtab file.

raiddev /dev/md0
raid-level  0
nr-raid-disks   5
persistent-superblock 1
chunk-size  1024

device  /dev/sdc1
raid-disk   0
device  /dev/sdd1
raid-disk   1
device  /dev/sde1
raid-disk   2
device  /dev/sdf1
raid-disk   3
device  /dev/sdb2
raid-disk   4

Note that I have tried several variations on this file including knocking the chunk 
size down to 4 and including the nr-spare-disks option.

Any help would be greatly appreciated.

Thanks,
Rob Jones

resource

[Robert Jones]

Is there a resource that provides info on the latest raidtools.  I can't
seem to get a raid 0 device up with the 0.90 tools.  I had no problems with
the 0.3d tools.

Thanks,
Rob Jones


--
Rob Jones, Network Administrator, Dublin, OH: 614-760-4018 ext. 64018;
Cell Phone: 614-440-1323; Pager: 1-800-759-8352 PIN #: 1582753
--

Re: RAID under 2.2.10

[Robert Stuart]

John E. Adams wrote:
> 
> >> Tom Livingston wrote:
> >>
> >>> As others have pointed out recently on this list, you can get raid working
> >>> with a 2.2.10 kernel.  Ingo posted a fix, which involves changing just one
> >>> line.
> >>
> >> I wrote:
> >> The fix is only one line, BUT that one line occurs TWICE.  Change both
> >> occurrences of 'current->priority = 0' to 'current->priority = 1'
> >> in /usr/src/linux/drivers/block/md.c.  Ideally, that constant should
> >> have a symbolic name like LOWEST_PRIORITY.
> >>
> > Christopher E. Browne wrote:
> >   So if I am distilling the correct data here, one patches 2.2.1
> > with the latest 2.2.6 raid patch, ignores the rejects, and cheges
> > those to lines and then has a working raid system?
> > Are there and issues with the AC patches?
> 
> Mostly correct.  The 2.2.6 patch fails against linux/include/linux/fs.h
> The following code, which is the failing piece, needs to be added to fs.h
> 
> static inline int buffer_lowprio(struct buffer_head * bh)
> {
> return test_bit(BH_LowPrio, &bh->b_state);
> }
> 
> I don't know about AC patches, I no longer apply them.
> 
> johna

I'm wanting to use the latest kernel with raid patches and I'm new to
the mailing list... Is raid with 2.2.10 a matter of applying the 2.2.6
raid patches, and adding that code above?  What are the "AC" patches?   
Is the fix in the second paragraph above required?  

What are good sites for raid info - can I find digests of this list
anywhere?  

Thanks for your help.

-- 
Robert Stuart
Ph  61-7-3864 0364

RAID 0 performance, SCSI & IDE

[Robert Ekl]

Hi.  I was wondering if I am getting degraded performance because I am
using both IDE & SCSI disks in my software raid0 array, over the
performance I would get if I used only SCSI disks.

Here's the setup:

md0 : active raid0 sdb1 sda1 hda7 3084032 blocks 256k chunks

/dev/hda:

 Model=WDC AC28400R, FwRev=17.01J17, SerialNo=WD-WM628
 Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs
FmtGapReq }
 RawCHS=16383/16/63, TrkSize=57600, SectSize=600, ECCbytes=40
 BuffType=3(DualPortCache), BuffSize=512kB, MaxMultSect=16, MultSect=off

 DblWordIO=no, maxPIO=2(fast), DMA=yes, maxDMA=0(slow)
 CurCHS=1027/255/63, CurSects=-66060037, LBA=yes, LBAsects=16514064
 tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2
 IORDY=on/off, tPIO={min:160,w/IORDY:120}, PIO modes: mode3 mode4
(5400 RPM, 512k buffer)

Vendor: Quantum   Model: VP32210   Rev: HP03
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sda at scsi0, channel 0, id 0, lun 0
 (5400 RPM, 1024k buffer,  Fast SCSI-2)

 Vendor: SEAGATE   Model: ST11200N ST31230  Rev: 0660
  Type:   Direct-Access  ANSI SCSI revision: 02
Detected scsi disk sdb at scsi0, channel 0, id 1, lun 0
(5400 RPM, 256k buffer, Fast SCSI-2)

scsi0: Target 0: Queue Depth 28, Synchronous at 10.0 MB/sec, offset 15
scsi0: Target 1: Queue Depth 28, Synchronous at 10.0 MB/sec, offset 15


Thanks.
Rob

Re: Patching kernel for RAID.

[Robert (Drew) Norman]

cat raid0145-19990421-2.2.6 | patch -p0

--Drew

On Fri, 11 Jun 1999, Mark Beck wrote:

> Guys,
> 
> Ive downloaded 2.2.6 kernel, and now trying to patch
> "raid0145-19990421-2.2.6".
> 
> [root@lisc0002 src]# ls -l
> total 394
> lrwxrwxrwx   1 root root   12 Jun 11 16:18 linux -> linux-2.2.6/
> drwxr-xr-x  15 1046 1046 1024 Jun 11 16:25 linux-2.2.6
> drwxr-xr-x  14 1046 1046 1024 Apr 28 16:49 linux-2.2.7
> drwxr-xr-x  15 1046 1046 1024 Jun  9 16:47 linux-2.2.9
> -rw-r--r--   1 root root   395628 Jun 11 16:21
> raid0145-19990421-2.2.6
> drwxr-xr-x   7 root root 1024 Jun  9 19:32 redhat
> 
> 
>  I now do:
> 
> [root@lisc0002 src]# patch  can't find file to patch at input line 3
> Perhaps you should have used the -p or --strip option?
> The text leading up to this was:
> --
> |--- linux/fs/block_dev.c.orig  Tue Nov 24 11:41:28 1998
> |+++ linux/fs/block_dev.c   Sun Apr 18 10:39:09 1999
> --
> File to patch:
> 
> What am I doing wrong ?
> 
> Thanks
> Mark
> 
> 

Re: Raid on Kernel 2.2.9

[Robert (Drew) Norman]

I have raid running with kernel 2.2.9-ac4.  I had to go about it in the
following manner:
untar kernel.   
Apply ac4 patch.  
Apply  raid0145-19990421-2.2.6
Apply the following patch for the failed patch of raid0145-1999 (I had to
manually apply it, it wouldn't install itself for some reason)


--- linux/include/linux/fs.h.orig   Sat May 15 11:10:15 1999
+++ linux/include/linux/fs.hSat May 15 11:11:21 1999
@@ -255,6 +255,11 @@
return test_bit(BH_Protected, &bh->b_state);
 }

+static inline int buffer_lowprio(struct buffer_head * bh)
+{
+   return test_bit(BH_LowPrio, &bh->b_state);
+}
+
 #define buffer_page(bh)(mem_map + MAP_NR((bh)->b_data))
 #define touch_buffer(bh)   set_bit(PG_referenced,
&buffer_page(bh)->flags)



--Drew Norman


On Fri, 11 Jun 1999, A James Lewis wrote:
> As far as I am aware there are no working patches for 2.2.8 on, this has
> been the case for some time so that information may be out of date... or
> alternatively there may be moves afoot to integrate it into a mainstream
> kernel.
> 
> Does anyone have clarification on this as I am installing some servers
> with 2.2.7 so as to get raid functionality
> 
> James
> 
> > Where can I find the Kernel patches to get Raid working correctly for
> > Kernel 2.2.9
> > 

Re: Success!

[Robert (Drew) Norman]

The best way to do a copy and make sure it all works is like this:

find . -print | cpio -pdmv 

--Drew Norman

On Thu, 10 Jun 1999, m. allan noah wrote:

> cp -a most assuredly works. gets me where i am going every time :)
> 
> allan
> 
> 
> Chris Jester <[EMAIL PROTECTED]> said: 
> 
> > I always use: cp -dpRv and it works just fine. :-)
> > 
> > 
> > 
> > At 4:32 PM -0700 6/10/99, David Robinson wrote:
> > >You did what?
> > >
> > >cp -a
> > >
> > >It gets them everytime:-)
> > >
> > >cp & symlinks are not friends...
> > >
> > >If you ever need to copy a directory or filesystem do a tar! It
> will not
> > >only keep the permissions & owner bits correct, It will also keep
> > >symlinks pointing the correct way..
> > >
> > >There is a trick to copy a directory from one place to another
> using tar
> > >but I can't remember it right now... something like:
> > >tar -cv . | (cd /newdir; tar -x --)
> > >I don't think that's it I'm sure someone else here knows it
> by
> > >heart:-)
> > >
> > >"Nixon, Shon" wrote:
> > >
> > >>
> > >>
> > >> I finally have both a RAID1 and RAID5 system up and running
> thanks to
> > >> the help of Mr. Allan Noah. I am however having trouble with
> symlinks
> > >> though. I tried to start X after I got the system up and
> running and
> > >> it would seem that xfs is having trouble finding fonts. I also
> tried
> > >> to compile dosemu to use X for xdos and it states that it can
> not find
> > >> the X11 libraries? I reinstalled X (several times) and it now
> works
> > >> (some functions are hairy though) but I still get the same
> message
> > >> with dosemu. My concern is that if X is having problems, then
> other
> > >> programs may also rear their ugly head with problems in the
> future. I
> > >> used cp -a copying each directory from root (/) one-by-one to a
> > >> /raidtest share (md0) that I created for the raid mount of
> root. I
> > >> also have /raidtest/usr (md2) ../home (md3) & ../var (md1) raid
> > >> mounts. Have I missed something here? This has happened with
> both
> > >> raid1 & 5.
> > >>
> > >> Shon Nixon
> > >> [EMAIL PROTECTED]
> > 
> > 
> > ** PLEASE NOTE ***
> > OUR NEW ADDRESS AND TELEPHONE NUMBER:
> > 
> > SplitInfinity
> > 13553 Poway Rd. #1504
> > Poway, CA 92064
> > U.S.A.
> > 
> > OUR NEW PHONE NUMBER: 619-679-2814
> > 
> > All items should be mailed to this address from this moment
> forward.
> > 
> > Thank you.
> > 
> > Chris Jester
> > SplitInfinity
> > 
> > 
> > 
> > 
> > 
> > 
> 
> 
> 

Re: raid 5 problems in 0.90 (fwd)

[Robert (Drew) Norman]

Just my $0.02.  I have been using IBM SCSI drives in linux software raids
for the past year.  I haven't seen any problems, and I think they are my
favorite drives.

But that might just be my luck.  I am running them in 3 or 4 different
raid arrays.

--Drew

On Mon, 31 May 1999, Francisco Jose Montilla wrote:

> On Mon, 31 May 1999, Till Mommsen wrote:
> 
>   Hi,
> 
> > >   I guess this is surely due to disks/controller (won't you be using
> > > IBM disks?) but certainly, you had luck in being able to build a raid with
> > > that config. Guess that the spare-disk wrong option went unadverted to the
> > > *picky* raid config file parser...
> > 
> > What is wrong with IBM Disks and Linux SW RAID?
> 
>   Nothing concrete, but I've seen lots of headaches with those
> drives in general. Gerard Roudier (ncr53c8xx driver) pointed out recently
> that these drives doesn't like the way some controllers arbitrate the
> SCSI bus...
> 
>   greetings,
> 
> *---(*)---**-->
> Francisco J. Montilla  Systems & Network administrator
> [EMAIL PROTECTED]  irc: pukkaSeville  Spain   
> INSFLUG (LiNUX) Coordinator. www.insflug.org   -   ftp.insflug.org
> 
> 

Re: Raid problems.

[Robert (Drew) Norman]

I would like to thank everyone for helping me solve my RAID problems.
With the kernel patch everything seems to work ok now.  I see great
increases in speed over the raid-0.50 I was running before.  I was
attempting to use a single disk as a raid because I was testing it before
i put it on a production machine.  I had no intentions of running raid on
a single disk.  However, the purpose was well served because I learned
what to do before I upgraded my existing file servers.  If I had not done
that, I would have broken my existing file server infrastructure
temporarily with the upgrades.  And downtime sucks.

--Drew


On Thu, 20 May 1999, Piete Brooks wrote:

> [ I've delayed reply to this one to try to work out how to express myself
>   better, but have failed, so sorry -- it's not as clear as I'd hoped
> ]
> 
> >> o Learning
> >>Granted, no speed improvements, but you can learn about it. 
> >>Knowledge like that comes in handy in Interviews ;-)
> > But you are learning in an environment not conducive to learning
> 
> I disagree.
> 
> > -> ie: the system is not real-world, and in fact, may be far from a good
> > example of the real world.
> 
> It's a learning curve.
> 
> Such testing gets you a long way up that curve !!
> 
> 1) have I correctly set all the kernel build options
> 2) do the startup scripts (etc) correctly start/stop the RAID
> 3) can I create syntactically correct raidtab files
> 4) can I invoke mkraid correctly
> 5) what are the procedures for fixing problems when a disk has failed
> 6) how can I add spare disks
> 7) how so I convert a raw disk to RAID1
> 8) What do I do when I just see "mkrai aborted"
> 
> If all users had performed all the above tests on such dummy systems,
> I suspect 90% of the please for help on this list would go !
> 
> > Again, testing in a environment not simulating real world.
> 
> Again, I diagree !
> 
> It *IS* using the same kernel, same scripts, same utilities, same procesures,
> etc, etc
> 
> > I can get a chair, and an old record (steering wheel) and pretend
> > that I am driving a car.
> 
> No -- a closer analogy is to get into a *REAL* car, and drive it round a
> deserted carparkl / runway / ...
> 
> You learn how to start the engine, control the clutch, stop the car,
> control a skid, etc.
> 
> Sure, you don't learn how to deal with other traffic, but you get used to
> driving a *REAL* car without the chance of doing damage to anyone else !
> 
> [[ I learnt to drive a car aged 10.
>I made one mistake, and knocked down our garden wall.
>I also learnt how to build a dry stone wall :-))
> ]]
> 
> > Sure, I can learn and 'test' things,
> 
> ... in a real car -- yup !
> 
> > but do they in any way represent a real world scenario?
> 
> It's a pretty close approximation of the basics.
> Lets get people a good half way up the learning curve before letting them loose
> on real data !!
> 

RE: RAID0 and RedHat 6.0

[Robert McPeak]

All -

Compiling non-modular support for RAID0 into the stock RedHat 6.0 kernel
(2.2.5-15) and adding the RAID0 partition to the fstab file solved the
automounting problem.  It seems there's a problem with either autofs or
raidtools here...  Thanks for the help!

Rob

-Original Message-
From: Ingo Molnar [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, May 18, 1999 5:05 AM
To: Robert McPeak
Cc: [EMAIL PROTECTED]
Subject: Re: RAID0 and RedHat 6.0



On Mon, 17 May 1999, Robert McPeak wrote:

> Here are the relevant messages from dmesg:

> hdd1's event counter: 000c
> hdb1's event counter: 000c
> request_module[md-personality-2]: Root fs not mounted
> do_md_run() returned -22

hm, this is the problem, it tries to load the RAID personality module but
cannot find it, because the root fs is not yet mounted. But
'md-personality-2' is strange as well, it should be 'md-personality-0' for
RAID0, there is no personality-2 ...

when you run it manually:

> raid0 personality registered

then it correctly registers raid0. You'll definitely get rid of these
problems if you compile RAID into the kernel (this is only a workaround),
but these things supposed to work. I'm not sure yet whats going on.

-- mingo

RAID0 and RedHat 6.0

[Robert McPeak]

Hi - 

 
I just 
installed RedHat 6.0, which appears to have the 0.90 version of the raidtools 
installed by default.  My boot disk is separate from the RAID.  I 
created a RAID0 spanning two 9gb drives, and it works fine, as long as I 
manually go in and to a raidstart and a mount after booting.  During the 
init scripts, it does try to start and mount the RAID, but fails.  How do I get this 
partition to autostart and mount?  Thanks!  

 
Rob 
McPeak (Linux Newbie)
 
 
Here 
are the relevant messages from dmesg:

autodetecting RAID arrays(read) hdb1's sb offset: 9873216 [events: 
000c](read) hdd1's sb offset: 9873216 [events: 000c]autorun 
...considering hdd1 ...  adding hdd1 ...  adding hdb1 
...created md0bindbindrunning: 
now!hdd1's event counter: 000chdb1's 
event counter: 000crequest_module[md-personality-2]: Root fs not 
mounteddo_md_run() returned 
-22unbindexport_rdev(hdd1)unbindexport_rdev(hdb1)md0 
stopped autorun DONE.

 
autorun ...considering hdd1 ...  adding hdd1 ...  
adding hdb1 ...created 
md0bindbindrunning: 
now!hdd1's event counter: 000chdb1's 
event counter: 000craid0 personality registeredmask 
fffc rdev->size: 9873216 masked rdev->size: 
9873216  new md_size: 9873216 rdev->size: 
9873216 masked rdev->size: 9873216  new md_size: 
19746432md0: max total readahead window set to 256kmd0: 2 data-disks, 
max readahead per data-disk: 128kraid0: looking at 
hdb1raid0:   comparing hdb1(9873216) with 
hdb1(9873216)raid0:   ENDraid0:   ==> 
UNIQUEraid0: 1 zonesraid0: looking at hdd1raid0:   
comparing hdd1(9873216) with hdb1(9873216)raid0:   EQUALraid0: 
FINAL 1 zoneszone 0 checking hdb1 ... contained as device 
0  (9873216) is smallest!. checking hdd1 ... contained as 
device 1 zone->nb_dev: 2, size: 19746432current zone offset: 
9873216done.raid0 : md_size is 19746432 blocks.raid0 : 
conf->smallest->size is 19746432 blocks.raid0 : nb_zone is 1.raid0 
: Allocating 8 bytes for hash.md: updating md0 RAID superblock on 
devicehdd1 [events: 000d](write) hdd1's sb offset: 9873216hdb1 
[events: 000d](write) hdb1's sb offset: 9873216 autorun 
DONE.

Raid problems.

[Robert (Drew) Norman]

I can not get raidtools-0.90 to work.  I have attempted everything I know
to do.  I have used the raidtools-0.50 before with no problems.  I am
running the following:

Redhat Linux 6.0 with 2.2.8 recompiled kernel.
448MB RAM
raidtools-0.90
I have a IBM 9GB drive split into 3 partitions of equal size.

Here is my raidtab:


raiddev /dev/md0
raid-level0
nr-raid-disks 3
nr-spare-disks0
chunk-size16

#persistent-superblock 0
device/dev/sdb1
raid-disk 0
device/dev/sdb2
raid-disk 1
device/dev/sdb3
raid-disk 2


[root@dnorman-pc /root]# mkraid -f /dev/md0
DESTROYING the contents of /dev/md0 in 5 seconds, Ctrl-C if unsure!
handling MD device /dev/md0
analyzing super-block
disk 0: /dev/sdb1, 2811343kB, raid superblock at 2811264kB
disk 1: /dev/sdb2, 2811375kB, raid superblock at 2811264kB
disk 2: /dev/sdb3, 2811375kB, raid superblock at 2811264kB
mkraid: aborted


Thanks in advance for the help.
--Drew Norman

Re: (Offtopic) Backup Systems

[Robert Siemer]

Re!

On Thu, 13 May 1999, Francisco Jose Montilla wrote:
>   My question is: what combination of their use do you think is
> better? more reliable? simpler? 

www.amanda.org

It is used here on the network. - I have no experience with the software,
but I will take it first, when I have a better backupsystem at all... [:

Bye,
Rob

One disk works more?

[Robert Siemer]

Hi!

> cat /proc/mdstat 
Personalities : [raid0] 
read_ahead 1024 sectors
md0 : active raid0 sdd1[1] sdc1[0] 17782528 blocks 16k chunks
unused devices: 

Yesterday:
sdc and sdd have only one partition, but the LED of sdd (and the
hostadapter) lits up every half second - even if nothing is done on md0!
sdd remains "off".
Does this make sense?

> uname -a
Linux panorama 2.2.2 #6 Fri Mar 5 01:47:41 CET 1999 i586 unknown


Bye,
Rob

Re: Server crash

[Robert Siemer]

Hi!

On Wed, 5 May 1999, Paul Hancock wrote:

> May  1 12:49:31 postal kernel: EXT2-fs warning (device md(9,0)):
> ext2_free_blocks: bit already cleared for block 73716
> May  2 10:23:18 postal kernel: EXT2-fs error (device md(9,0)):
> ext2_check_descriptors: Block bitmap for group 8 not in group (block 0)!
> May  2 10:23:18 postal kernel: EXT2-fs: group descriptors corrupted !

> problem could be.  I upgraded to 2.2.6 after my e-mail server crashed a
> couple of weeks ago.  It was running raid 1 + 2.1.115, and had run for
> several months without a problem before that.

This is usualy not a problem in ext2-fs. - And I think in raid, neither.

Something in 2.2.x was improved drastically, that my SCSI-adapter made
problems. I changed the external clock from 83 MHz back to 66 MHz and the
PCI-bus from async. 32 MHz to sync. 33 MHz.

Before that I could reproduce a system-stop with:

ls -flaR / &
cat /dev/sda > /dev/null &
cat /dev/sdb > /dev/null &
cat /dev/sdc > /dev/null &
cat /dev/sdd > /dev/null &

The update-daemon was also running. If this "heavy" SCSI-load corrupts
something, there could be a problem with the hardware.

Or maybe I'm completely wrong...  (:

Bye,
Rob

PS: Read also linux-kernel-mailinglist-FAQ...

Watchdog-problem [was: Raid crash and burn issues.]

[Robert Siemer]

Hi!

On Tue, 27 Apr 1999, Alvin Starr wrote:

> I tried to use the software watchdog on a system running with a raid1 root
>  [...]
> At that point I was hung and had to perform a hard
> reset.

What is "cat /proc/sys/kernel/panic" saying?


Bye,
Rob

Re: Help with mirroring

[Robert Siemer]

Hi!

On Tue, 20 Apr 1999, Jason Speckman wrote:

>  I have kernel 2.2.5 with raid 1 compiled in, raidtools 0.90.  I
> want to mirror my /dev/sda2 and /dev/sdb2 partitions.  I already have
> (problems.) [...]

Have you patched the kernel? A patch is needed for raidtools. Check the
FAQ and HOWTO? (Is it explained there?)

Patch is available under
ftp://ftp.iso.kernel.org/pub/linux/daemons/raid/ALPHA/raid0145...
or slightly different...

Bye,
Rob

Re: Swap on raid

[Robert Siemer]

Hi!

On 15 Apr 1999 [EMAIL PROTECTED] wrote:

> AFAIK, the swap code uses raw file blocks on disk, rather than passing
> through to vfs, cause you dont want to cache swap accesses, think about
> it :) 
> 
> this is how swap can work on a partition or a file, cause at swapon
> time, the blocks are mapped for direct access. 

I'm currently using NFS-Swap with kernel 2.0.3x. - On a 10MB/s-net...
This was a "standard-patch" (from bigmama?).
Would this work with a local nfsd with raid?
Sure - this has nothing to do with performance...  [:

Okay, okay, this is completely pointless.   (:

Bye,
Rob

Re: Filesystem corruption (was: Re: Linux 2.2.4 & RAID - success report)

[Robert Siemer]

Hi!

On Fri, 9 Apr 1999, Brian Leeper wrote:

> On Wed, 31 Mar 1999, Tony Wildish wrote:
> >  Try booting with the 'mem=xxxM' option to limit yourself to a small
> > amount of RAM. If you are lucky then the problem is high enough in the
> > memory that you can limit yourself to a good region and it will work.
> 
> There's also a memtest86 utility that compiles under Linux to produce a
> disk you can boot your system with and test the memory--I've find some bad
> memory with that before.
> 
> Compiling a kernel a bunch of times is also a good way to test if you've
> got bad memory.

Before I knew about memtest I compiled the kernel several times. It failed
after one loop. The second time it failed after the 50th loop (2 days
running). I checked nearly everything, no solution. (I got a stable
"snapshot of my configuration".)
After one year the problem occoured again on this machine. Now I know the 
problem. It was the ADVANCED POWER MANAGEMENT. Switch it off. Always. And
read the tips in the kernel-configuration-help to APM in 2.2.x.

Bye,
Rob

Re: Low performance on K6

[Robert Siemer]

Re!

On Wed, 31 Mar 1999 [EMAIL PROTECTED] wrote:

> hardware:
> 2 x IBM UltraStar 9ES 4.5 GB U2W
> Adaptec 2940 UW
> AMD K6 - 200 MHz
> 64 MB RAM
> 
> Each disk gives us around 11 MB/s using hdparm -t /dev/sd[ab]
> 
> If we run  hdparm -t /dev/sda & hdparm -t /dev/sdb,  (two hdparm's
> concurrently) we get roughly 8.5 MB/s from each disk,  a total of
> 17 MB/s.

Is my assumtion correct, that hdparm -t uses the disks nearly linear?
My two disk give with
dd if=/dev/sdc of=/dev/null  or  dd if=/dev/md0 of=/dev/null
about 10 MB/s and together in the md0-Device abount 20 MB/s.

> However, when we set up /dev/md0,  a RAID-0 device between the two
> disks, (in a configuration similar to what I've run myself),  we get
> around 10 MB/s from the md0 device. --   Less that we can get 
> from one single disk.

> Does anyone have a clue about what could be going on ?? The main 
> difference between my own box which ran a similar configuration for long
> time seems to be that this box is a K6.Anyone running raid0 on similar 
> configurations with a K6 CPU ?

I'm using a K6-2 333Mhz.


Bye,
Rob

Re: RAID Docs

[Robert Siemer]

Re!

On Thu, 18 Mar 1999, Rolando Lopez wrote:
> My name is Rolando and I downloaded raidtools-0.90 RPM from
> 
>http://niteowl.userfriendly.net/linux/RPM/contrib/libc6/i386/raidtools-0.90-19990309.i386.html

RPM? This shows again, that looking for *.tar.gz is the ultimate
solution... (:

> I'm wondering what I'm doing wrong. I run a 2.2.2 kernel with RAID0
> [...]
> # raidstart -a
> /dev/md0: Invalid argument

Yes, I tried weeks, too. What you need with kernel 2.2.x:

- raid0145 kernel-patches
- corresponding raidtools

There is also a new HOWTO in the tar-ball. I've never read the new one.
You get the files per ftp:
ftp://ftp.xx.kernel.org/pub/linux/daemons/raid/ALPHA/... (or slightly
different)

Only the patches for 2.2.3 can be applied without human interaction. 
So upgrade to 2.2.3...

Bye,
Rob

Please help with OLD Raid signature....

[Robert O'Kane]

Hello,

  OK, OK, I know.. it is all pre-beta software but I need help
with it now. We lost some 16Gb now due to some stupid drives
and drivers.

We have two md-raid 5 arrays running on an 'older' version of the
md-tools. -- /usr/src/linux/include/linux/md.h says 0.36.5
raidadd v0.3d compiled for md raidtools-0.50

Linux deepthought 2.0.35 #5 Fri Jul 24 17:05:34 CEST 1998 i686
unknown

What happened is there was a dead drive in the array, then the
scsi-bus
got reset (stupid aha152x somehow reset the WHOLE scsi sub-system
including
the 2 ncr810a boards for the raid arrays!  AAARRRGGG)

Anyway, the strange part SEEMS to be that the MD tried to
rebuild,
and for some unknown reason, a second drive on the failed array
also failed temporarily. MD couldn't 'update' the timestamp on
the
md-header and the whole thing just crapped out but not before
writting
the update to the other 2 drives in the array. Now, I'm 97.07734%
sure
the temporary faild disk has good data on it as the whole array
just
stopped at that point. Its just missing the timestamp.

HOW can I create a timestamp on the 'not so failed' disk so I can
get
the (some?) data off?
(My DLT backup is 2 months old and we need the data!)

I promise to upgrade the machine's raid system after I get this
data!

Thanks for the help and advice!

Bob.

*/
------
WHO   : Robert O'Kane
WHERE : Künstlergruppe OTHERSPACE
WHAT  : Founding Member and Linker

WHY   : http://www.rz.uni-frankfurt.de/~okane
HOW   : [EMAIL PROTECTED]
--

mkraid aborted without comment [was: advice]

[Robert Siemer]

Hi!

On Tue, 9 Mar 1999, Mike Dickson wrote:
> I did:
> -compiled kernel 2.2.3 with MD support built in for all levels (not as modules)
> -installed raidtools-0.90-19990128.i386.rpm
> -created /etc/raidtab:
> #start raidtab
> raiddev /dev/mdo
^^^  /dev/md0 is better

>   raid-level  0
>   nr-raid-disks   2
>   nr-spare-disks  0
> 
>   device  /dev/sdb1
>   raid-disk   0
>   device  /dev/sdc1
>   raid-disk   1
> #end raidtab
> -issued:
> % mke2fs /dev/sdb1 -R stride=64
> % mke2fs /dev/sdc1 -R stride=64

You dont need this. The ext2-FS should be on /dev/md0, the RAID0-Device.

> % mkraid -f /dev/md0 
>   DESTROYING the contents of /dev/md0 in 5 seconds, Ctrl-C is unsure!
>   handling md device /dev/md0
>   analyzing super-block
>   disk 0: /dev/sdb1, 4441941kB, raid superblock at 4441856kB
>   disk 1: /dec/sdc1, 8883913kB, raid superblock at 8883840kB
>   mkraid: aborted

Is the raid01445-19990309-2.2.3.gz patch applied?
- Anyway, the HOWTO in raidtools-19990309-0.90.tar.gz is the most
uptodate - I think so.

Bye,
Robert

Re: a newbie loser

[Robert Siemer]

Re!

On Tue, 9 Mar 1999, A James Lewis wrote:
> I think there may be a problem with your mail.  I see no content.

> Current time: 1999/03/09 12:51 GMT
> ERROR: Header line added to ASCII armor: "Hash: SHA1"
> ASCII armor corrupted.

Your PGP has problems to handle this PGP-6.x signed mail (my too). Just
look directly in your Mailbox as a temporarly solution ("less ~/mailbox").

Bye,
Robert

Re: dev/md0: Invalid argument (?!)

[Robert Siemer]

Re!

On Thu, 25 Feb 1999, emile wrote:

> I'm wondering what I'm doing wrong. I run a 2.2.2 kernel and try to use
> the 0.90 raidtools but when I'm trying something like:
> # raidstart /dev/md0
> I get:
> /dev/md0: Invalid argument
> [...]

> the tarball I used is
> raidtools-19990128-0.90.tar.gz and the linux kernel is 2.2.2 with md and
> RAID1 enabled.

> I've been unsuccesful in finding related info on this problem.

Yeah! That's really true. Documentation for the 2.2-kernel is not very
clear, even in the howto! 
I thought that enableing raid in the kernel-config is enought, too. But in
reality you need to apply also the raidpatch (raid0145-*2.2.0*[gz|bz2]
somewhere by ftp.xx.kernel.org/pub/*/daemons/raid/alpha/). The patch works
at least with 2.2.1, too. Some patches will be rejected (mmap.c, genhd.c),
here you need only to modify the *.c on your own, this should be easy,
because only the lines differs and things which are going to be removed
are not completely the same.

I will try this today with 2.2.2 (already patched, but not compiled...).


Does it work now?
Bye,
Robert

Re: was: swapping to file doesn't work: swapon hangs ....

[Robert Siemer]

Re!

On Thu, 25 Feb 1999, Benno Senoner wrote:

> Some time ago I patched an old 2.0.12 kernel with the swap-over-NFS
> patch, for a diskless
> machine and it worked very well.

Aehm, is there a patch for 2.2.x to be able to swap over NFS?! I need this
really for my diskless machine...

> Does anyone know if there is a similar pacht that allows to swap over a
> soft RAID5 disk ?

Here on a machine with kernel 2.0.35 I use the following in my /etc/fstab:
/dev/md0swapswapdefaults 0 0

Okay, md0 is a raid0-array, and this could be done without raid (similar
priorities for usual swap-partitions), but why not with raid5?

Or is the intention only to swap in a file...


Bye,
Robert

Re: Linux Ping problem

[Robert Siemer]

Re!

On Wed, 24 Feb 1999 [EMAIL PROTECTED] wrote:
> Previously before inserting my Linux hostname to DNS server, I could run
> ping program.

Has this something to do with RAID?? 

> After inserting, I can run ping. It says "Ping : ping must run as root"
> I checked the access control. It shown "-rwxr-xr-x   1 root senoko
> 14340 May  6  1998 ping"

Wrong. Do "chmod 4755 ping" as root.

> The other problem is when I log-in as root, I can't ping to myself (either
> using IP address or DNS names) but I can ping other systems. Similarly,
> from other systems I cang ping my Linux using IP and DNS names.

If DNS-resolving works, IP and the DNS-Name are nearly the same... Maybe
your're using the wrong IP/Name at all. Look at the output of "ifconfig".
Check also "ping 127.0.0.1" and ask the next nearby linux-profi in your
town or look at www.linux.org for a mailinglist, which fits better that
linux-raid. (!)

Bye,
Robert

RE: Compaq Smart Array & RH6.1

[Gerrish, Robert]

> From: Roeland M.J. Meyer on Friday, November 19, 1999 10:05 AM
> For smilar price, one can buy ready-built Linux boxen from
> . But, I've had excellent results 
> assembling my own servers as I need them. Most of them are
> standardized on ATX cases, Asus motherboards, and Adaptec SCSI cards.

That's exactly what we have been doing here, same formula, ATX /
Asus / adaptec with Seagate drives.  You know what you are getting
and you can control the quality of the parts.

Bob Gerrish

RE: Compaq Smart Array & RH6.1

[Gerrish, Robert]

> From: Matthew Clark 
> Sent: Thursday, November 18, 1999 10:21 AM
> 
> We are thinking about purchasing a Compaq Proliant 5500 - I 
> notice you were
> both experiencing problems with the Smart RAID controller 
> under Linux.  Can
> I ask if these problems have been resolved and if you are 
> getting decent
> performance from the RAID arrays?

Installation was a problem and I haven't been able to upgrade.
I will need to compile custom kernels, custom boot/install disks
and all that if I want to risk upgrading the now running system.

> Any information would be appreciated! (I just wasted a lot of 
> time working
> out that the I/O performance on the Hewlett packard LH4 is 
> unacceptable)

We have a whole new IT department in here since Jan 1st, and 
between three of us we have 20 years repair / support experience.
We have all had our fill of Compaq (and other priprietary computers.)
We have another Proliant that we just retired as a novell server.
One of the disks in the Smart 2 array failed and we replaced it 
with an identical disk.  We can't configure the array and have just 
retired the machine instead of turning into a linux or NT server.
Three of us have 20 years repair experience and

Do yourself a favor and get a generic off the shelf PC with high
quality parts.  You an easily repair and upgrade it.

RE: errors on boot

[Robert Dahlem]

Michel,

On Thu, 18 Nov 1999 11:00:36 -0500, Michel Pelletier wrote:

>> >> > physical?  The other two partions on the same drive work 
>> dandy with
>> >> > their RAID, it looks like just this one partition on the 
>> >> drive is bad.
>> >> > How, in a nutshell, do I fix this sucker?
>> >> > 
>> >> try raidhotremove /dev/mdx /dev/sdy
>> >> raidhotadd /dev/mdx /dev/sdy
>> >
>> >Hmm... when I did:
>> >
>> >raidhotremove /dev/md2 /dev/sda
>> >
>> >It told me that /dev/sda was not in the array, when it is, 
>> /dev/md2 is
>> >built from /dev/sda4 and /dev/sdb4.
>> 
>> Well, you should not try to remove /dev/sda as its not part 
>> of /dev/md2.
>> 
>> Try /dev/sda4 ...
>
>I tried that also.  In fact, I tried all combinations, sorry I wasn't
>explicit about that.  I get an error that /dev/sdb4 is not in the array,
>and /dev/sda4 is busy.
>
>I'm not even sure which one is the good one!  I'm assuming that since
>/proc/mdstat tells me the second on is bad, and that /dev/sdb4 is the
>*second* defined drive for md2 in /etc/raidtab, that /dev/sdb4 is the
>faulty one (the log messages also point to this).

You can tell exactly by inspecting /proc/mdstat and /etc/raidtab

In /etc/raidtab (if not modified after mkraid) you will find something like

device  /dev/sda4
raid-disk   0
device  /dev/sdb4
raid-disk   1

When you find [U_] then the U is for /dev/sda4 (raid-disk 0) and the underscore is for 
/dev/sdb4.

>From what I understand from your mails you once lost one of your partitions 
>(/dev/sdb4). 
I think you should not try to raidhotremove it if its not in /proc/mdstat.

If you think the disk and the partition ist physically ok, you can try to get back to 
some stable situation. The following is the save way, but will eventually give some 
warnings.

raidsetfaulty /dev/md0 /dev/sdb4
(will eventually complain if partition already removed)

Now /dev/sdb4 ist still in /proc/mdstat (if it was before) but marked with an 
underscore 
and marked (F).

raidhotremove /dev/md0 /dev/sdb4
(will complain with "disk not in array" if already removed)

Now /dev/sdb4 should have vanished from /proc/mdstat

raidhotadd /dev/md0 /dev/sdb4

back in /proc/mdstat, still with underscore, but recovery running.

If this doesnt work, perhaps you should show us your /etc/raidtab and /proc/mdstat.

Regards,
Robert

RE: errors on boot

[Robert Dahlem]

Michel,

On Thu, 18 Nov 1999 10:24:15 -0500, Michel Pelletier wrote:

>> > physical?  The other two partions on the same drive work dandy with
>> > their RAID, it looks like just this one partition on the 
>> drive is bad.
>> > How, in a nutshell, do I fix this sucker?
>> > 
>> try raidhotremove /dev/mdx /dev/sdy
>> raidhotadd /dev/mdx /dev/sdy
>
>Hmm... when I did:
>
>raidhotremove /dev/md2 /dev/sda
>
>It told me that /dev/sda was not in the array, when it is, /dev/md2 is
>built from /dev/sda4 and /dev/sdb4.

Well, you should not try to remove /dev/sda as its not part of /dev/md2.

Try /dev/sda4 ...

Regards,
Robert

raidtools-0.50beta

[Robert B.]

I downloaded raidtool's version .5 beta and quit start raid guide and the
how to's are very outdated and do nothing but confuse the readers.  I have
tried to install them by following the direction and have been unsuccessful
is there any chance of an updated QuickStart.RAID or How to's? and I went to
Miguel's site @ luthien.nuclecu.unam.mx/~miguel/raid and all I found in his
directory was a GIF of some software probably gtk+ compliant I might add.
If your out there please help! Thanks in advance!

Robert B.